DNA Repair Factor RAD18 and DNA Polymerase Polκ Confer Tolerance of Oncogenic DNA Replication Stress

Published Online: 23 August, 2017 | Supp Info: http://doi.org/10.1083/jcb.201702006 JCB: Article Downloaded from jcb.rupress.org on November 1, 2018

DNA repair factor RAD18 and DNA polymerase Polκ confer tolerance of oncogenic DNA replication stress

Yang Yang,1 Yanzhe Gao,1 Liz Mutter‑Rottmayer,1 Anastasia Zlatanou,1 Michael Durando,1 Weimin Ding,1,4 David Wyatt,2,3 Dale Ramsden,2,3 Yuki Tanoue,5 Satoshi Tateishi,5 and Cyrus Vaziri1,2

1Department of Pathology and Laboratory Medicine, 2Lineberger Comprehensive Cancer Center, Curriculumin Genetics and Molecular Biology, and 3Department of Biochemistry and Biophysics, University of North Carolina at Chapel Hill, Chapel Hill, NC 4Oncology Center, Zhujiang Hospital, Southern Medical University, Guangzhou, China 5Division of Cell Maintenance, Institute of Molecular Embryology and Genetics, Kumamoto University, Kumamoto, Japan

The mechanisms by which neoplastic cells tolerate oncogene-induced DNA replication stress are poorly understood. Cyclin-dependent kinase 2 (CDK2) is a major mediator of oncogenic DNA replication stress. In this study, we show that CDK2-inducing stimuli (including Cyclin E overexpression, oncogenic RAS, and WEE1 inhibition) activate the DNA repair protein RAD18. CDK2-induced RAD18 activation required initiation of DNA synthesis and was repressed by p53. RAD18 and its effector, DNA polymerase κ (Polκ), sustained ongoing DNA synthesis in cells harboring elevated CDK2 activity. RAD18-deficient cells aberrantly accumulated single-stranded DNA (ssDNA) after CDK2 activation. In RAD18-depleted cells, the G2/M checkpoint was necessary to prevent mitotic entry with persistent ssDNA. Rad18−/− and Polκ−/− cells were highly sensitive to the WEE1 inhibitor MK-1775 (which simultaneously activates CDK2 and abrogates the G2/M checkpoint). Collectively, our results show that the RAD18–Polκ signaling axis allows tolerance of CDK2-mediated oncogenic stress and may allow neoplastic cells to breach tumorigenic barriers.

Introduction

During tumorigenesis, neoplastic cells must endure DNA dam- tail collisions between replication forks produce double-strand age from environmental, metabolic, and other intrinsic sources breaks (DSBs; Davidson et al., 2006). It is unknown whether (Bartkova et al., 2006; Halazonetis et al., 2008). Oncogene- oncogene-induced rereplication is caused by inappropriate induced DNA replication stress can be a major cause of intrinsic activation of DNA replication licensing factors, initiation fac- DNA damage and represents a potential source of genome in- tors, or deregulation of both licensing and initiation phases of stability in cancer cells. Many oncogenes, including v-RAS, cy- DNA synthesis. It is also unclear whether common mechanisms clin E, and others, induce DNA replication defects that trigger mediate rereplication and DNA damage in response to all on- DNA damage signaling (including ATM–CHK2, ATR–CHK1, cogenes. It is possible that the constitutive mitogenic signals and p53) and lead to irreversible cell cycle exit often termed induced by oncogenes culminate in aberrant cyclin-dependent oncogene-induced senescence (OIS; Bartkova et al., 2006; kinase 2 (CDK2) activation, in turn leading to DNA rereplica- THE JOURNAL OF CELL BIOLOGY CELL OF JOURNAL THE Di Micco et al., 2006). tion and other replication defects. Indeed, oncogene-induced The precise mechanisms by which oncogenes induce DNA replication stress is often modeled experimentally by DNA damage are incompletely understood. Oncogene-induced overexpression of CDK2 activators (Cyclin E and CDC25A) or DNA damage has been attributed to induction of genotoxic re- inhibition of the WEE1 kinase to remove negative constraints active oxygen species (ROS; DeNicola et al., 2011), depletion over CDK2 (Sogo et al., 2002; Bartkova et al., 2006; Beck et of nucleotide pools (Bester et al., 2011), collisions between the al., 2010, 2012; Jones et al., 2013). DNA replication and transcriptional machinery (Jones et al., Despite our limited mechanistic understanding of how 2013), or aberrant reinitiation of DNA synthesis multiple times oncogenes dysregulate DNA synthesis and cause DNA dam- each per cell cycle—a process usually termed “rereplication” age, there is general consensus that OIS poses a barrier to or “hyperreplication” (Di Micco et al., 2006). Rereplication tumorigenesis. Clearly, however, the OIS barrier is imperfect likely generates “onion skin” DNA structures in which head-to- and can be breached. The precise mechanisms by which oncogene-expressing cells withstand replication stress and DNA Correspondence to Cyrus Vaziri: [email protected] damage are poorly understood. DNA repair and/or DNA dam- Abbreviations used: DSB, double-strand break; HR, homologous recombination; IP, immunoprecipitation; MEF, mouse embryonic fibroblast; NHF, normal human © 2017 Yang et al. This article is distributed under the terms of an Attribution–Noncommercial– fibroblast; OIS, oncogene-induced senescence; PCC, premature chromatin con- Share Alike–No Mirror Sites license for the first six months after the publication date (see http densation; PCNA, proliferating cell nuclear antigen; PDS, pyridostatin; ROS, ://www.rupress.org/terms/). After six months it is available under a Creative Commons reactive oxygen species; sgRNA, single guide RNA; ssDNA, single-stranded License (Attribution–Noncommercial–Share Alike 4.0 International license, as described at DNA; TLS, trans-lesion synthesis; TMEJ, θ-mediated end joining; XPV, XP-variant. https://creativecommons.org/licenses/by-nc-sa/4.0/).

The Rockefeller University Press J. Cell Biol. Vol. 216 No. 10 3097–3115 https://doi.org/10.1083/jcb.201702006 JCB 3097 age tolerance capacity could potentially impact whether DNA In addition to its central role in TLS, RAD18 contrib- synthesis and viability are sustained when cells experience on- utes to activation of other genome maintenance pathways, cogenic stress. Interestingly, the DNA polymerase δ subunits including the Fanconi Anemia pathway (Palle and Vaziri, POLD3 and POLD4 can facilitate DNA replication in cyclin 2011; Williams et al., 2011), interstrand cross-link repair E–overexpressing cells (Costantino et al., 2014). Moreover, (Räschle et al., 2015), and DSB repair (Huang et al., 2009). the ATR–CHK1 pathway can promote oncogene-induced car- As the apical component of the TLS polymerase switch and cinogenesis (Schoppy et al., 2012). Therefore, DNA damage other DNA repair pathways, RAD18 expression levels and signaling and genome maintenance might critically influence activity critically impact DNA damage sensitivity. In neo- whether oncogene-expressing cells breach the OIS barrier. plastic cells, pathological activation of RAD18 by its stabi- However, there has been no systematic analysis of how DNA lizing binding partner MAGE-A4 has the potential to confer damage signaling and repair mechanisms impact DNA replica- both DNA damage tolerance and mutability, two important tion and cell cycle progression of oncogene-expressing cells. hallmarks and enabling characteristics of tumor cells. Be- It remains to be investigated whether all genome maintenance cause oncogenic stress is a major source of intrinsic DNA mechanisms or only specific subpathways of the DNA damage damage experienced by cancer cells, we hypothesized a role response confer oncogenic stress tolerance. Importantly, many for the RAD18 pathway in the tolerance of oncogene-induced cancer chemotherapeutic agents act by causing DNA repli- DNA damage. We show in this study that RAD18 is acti- cation stress and DNA damage. The selective pressures for vated in response to oncogenic stimuli that induce aberrant preneoplastic cells to acquire DNA damage tolerance during CDK2 activity. Moreover, RAD18 and its effector Y family tumorigenesis could also provide a mechanism for chemore- DNA polymerase Polκ (but surprisingly not the default TLS sistance. Therefore, the mechanisms by which cancer cells tol- polymerase Polη) are necessary for S phase progression and erate oncogenic DNA replication stress represent therapeutic viability in the face of excess CDK2 activity. Collectively, targets whose inhibition could sensitize tumors to intrinsic and this work suggests that TLS may be an important driver of therapy-induced DNA damage. carcinogenesis that promotes both tolerance of oncogenic We recently found that many cancer cells co-opt an ab- stress and mutagenesis. errantly expressed meiotic protein, the cancer/testes antigen MAGE-A4, to pathologically activate trans-lesion synthesis (TLS; Gao et al., 2016a). Cancer cell–specific RAD18 pathway Results activation by MAGE-A4 first suggested to us a possible role for TLS in the tolerance of replicative stresses that are unique Acute oncogene expression promotes to neoplastic cells. TLS is a specialized mode of DNA repli- RAD18-mediated PCNA monoubiquitination cation involving the DNA damage–tolerant and error-prone in untransformed cells Y family DNA polymerases η (Polη), κ (Polκ), and ι (Polι) We determined whether the TLS branch of the DNA damage re- as well as REV1 (Prakash et al., 2005). Individual TLS poly- sponse is activated in response to acute oncogene expression in merases perform replicative bypass of preferred cognate DNA human cells. We selected representative stimuli that are known lesions. Collectively, the TLS polymerases support DNA rep- to induce DNA replication stress and OIS, including Ha-RASV12, lication and viability in cells harboring damaged genomes. Y Cyclin E, c-MYC, and CDT1 (Bartkova et al., 2006; Liontos et family polymerase-deficient cells are often sensitive to agents al., 2007; Srinivasan et al., 2013). Oncogenes were transiently that cause replication fork–stalling and DNA replication stress, expressed in untransformed normal human fibroblasts (NHFs) demonstrating a key role for the TLS pathway in DNA dam- using adenovirus vectors. 48 h after infection, overexpressed age tolerance. However, owing to the low fidelity of Y family Cyclin E, C-MYC, and CDT1 proteins were readily detectable polymerases, TLS can be mutagenic (Prakash et al., 2005). by immunoblotting, and MAPK phosphorylation was used to Thus, TLS must be regulated stringently and used sparingly to validate RAS pathway activation (Fig. 1 A). maintain genome stability. As expected from previous work, DNA damage markers The E3 ubiquitin ligase RAD18 is a proximal activa- including CHK1 pS317 and γH2AX were induced by ectopi- tor of all four Y family TLS DNA polymerases (Hedglin and cally expressed Cyclin E and RAS (Fig. 1 A; Bartkova et al., Benkovic, 2015). In response to DNA replication fork stalling, 2006; Di Micco et al., 2006; Neelsen et al., 2013a). PCNA RAD18 monoubiquitinates proliferating cell nuclear antigen monoubiquitination was not significantly affected by MYC (PCNA) at the conserved residue K164. Y family DNA poly- or CDT1. However, levels of monoubiquitinated PCNA were merases possess ubiquitin-binding (ubiquitin-binding motif increased 25-fold in response to overexpressed Ha-RAS and [UBM] and ubiquitin-binding zinc finger [UBZ]) domains and 46-fold after Cyclin E expression. The fold changes in PCNA PCNA-interacting peptide boxes that facilitate their preferential monoubiquitination after Ha-RAS and Cyclin E expression association with PCNA in its monoubiquitinated form (Bienko substantially exceeded the threefold increase in PCNA modi- et al., 2005). The DNA damage–inducible interaction between fication induced by UV irradiation in NHFs (Fig. 1 A, second TLS polymerases and PCNA is the basis for the polymerase lane from left). The levels of ectopically expressed cyclin E switch that replaces replicative DNA polymerases ε and δ with that induced PCNA monoubiquitination in untransformed NHF Y family polymerases at stalled DNA replication forks. Polη cells were similar to steady-state Cyclin E levels in various exists in complex with RAD18 and is probably the default Y cancer cell lines including HCT116 colon cancer cells, A549 family TLS polymerase recruited to stalled DNA replication and H1299 lung carcinoma cells, and HeLa cervical cancer forks (Watanabe et al., 2004; Day et al., 2010). Polη is versa- cells (Fig. 1 B). In immunoprecipitation (IP) kinase assays, the tile and can perform replicative bypass of many bulky lesions, CDK2 activity of Cyclin E–overexpressing cells was compa- possibly explaining why this TLS polymerase is preferentially rable to CDK2 activity in the representative lung cancer cell recruited to all stalled replication forks. line H1299 (Fig. 1 C). Therefore, our standard Cyclin E over-

3098 JCB • Volume 216 • Number 10 • 2017 Figure 1. Oncogenic stimuli promote RAD18-mediated PCNA monoubiquitination. (A) Adenoviral vectors were used to express c-MYC, Ha-RASV12, CDT1, and Cyclin E in cultured NHF. After 48 h, cell lysates were analyzed by SDS-PAGE and immunoblotted with the indicated antibodies. (B) Immunoblot showing relative Cyclin E protein levels in AdCyclin E–infected NHFs and various cancer cell lines. Soluble extracts from the indicated cells were normalized

RAD18 confers oncogenic stress tolerance • Yang et al. 3099 expression conditions are appropriate for modeling the aberrant including mouse embryonic fibroblasts (MEFs) and U2OS Cyclin E–CDK2 signaling of cancer cells. cells (as shown in Figs. 1 F and 4). In RAS-transformed BJ human fibroblasts (Hahn et al., PCNA is a substrate for monoubiquitination by sev- 1999), CDK2 activity was approximately fourfold higher than eral E3 ligases including RAD18 (Watanabe et al., 2004), in isogenic parental BJ fibroblasts (Fig. 1 C) and was compa- CRL4Cdt2 (Terai et al., 2010), and HLTF (Lin et al., 2011). rable to CDK2 activity of AdCyclin E–transduced fibroblasts However, Cyclin E overexpression induced the formation of and H1299 lung cancer cells (Fig. 1 C). PCNA was constitu- RAD18 nuclear foci (Fig. 1 G). Moreover, MK-1775–induced tively monoubiquitinated in BJ-RAS cells when compared PCNA monoubiquitination was attenuated in RAD18-depleted with parental BJ fibroblasts (Fig. 1 D). Moreover, acute treat- NHFs and in Rad18−/− MEFs relative to isogenic RAD18- ment with the CDK2 inhibitor roscovitine attenuated PCNA replete cells (Fig. 1, E and F). Ectopically expressed siRNA- monoubiquitination in BJ-RAS cells (Fig. 1 D). We conclude resistant RAD18 fully rescued the PCNA monoubiquitination that Cyclin E– or RAS-induced CDK2 activity stimulates defect of RAD18-depleted cells (Fig. 1 H). Therefore, we con- PCNA monoubiquitination. clude that PCNA monoubiquitination in response to elevated As an additional approach to test the effect of CDK2 acti- CDK2 is RAD18 mediated. vation on PCNA monoubiquitination, we used a pharmacological inhibitor of WEE1 kinase (MK-1775) to de-repress CDK2. Cyclin E–CDK2–induced PCNA Using IP–kinase assays, we confirmed that CDK2 was activated monoubiquitination requires replication in MK-1775–treated cells (Fig. 1 C) as expected from previous licensing and origin firing work (Beck et al., 2012; Sørensen and Syljuåsen, 2012; Sakuri- Cyclin E–CDK2 is active in both G1 and S phase, and kar et al., 2016). Similar to ectopic Cyclin E or RAS expres- RAD18-mediated PCNA monoubiquitination can occur sion, MK-1775 treatment induced PCNA monoubiquitination throughout the cell cycle (Daigaku et al., 2010; Zlatanou et al., in NHFs (Fig. 1 E). The MK-1775–induced increases in CDK2 2011; Diamant et al., 2012; Yang et al., 2013). Therefore, we activity and PCNA monoubiquitination (sixfold and 4.3-fold investigated the cell cycle stage specificity and proximal trigger increases, respectively) were modest compared with AdCy- of PCNA monoubiquitination in response to aberrant Cyclin E– clin E–induced CDK activity and PCNA monoubiquitination, CDK2 activity. First, we asked whether PCNA monoubiquitina- indicating that MK-1775–induced CDK2 activity is limited by tion was associated with the imbalance of licensing or initiation endogenous Cyclin E levels. phases of DNA replication. We ectopically overexpressed DNA Because MK-1775 de-represses both CDK1 and CDK2, replication licensing factors (CDT1 and CDC6) or Cyclin E to we sought to determine which CDK contributes to PCNA stimulate replication licensing or initiation of DNA synthesis, monoubiquitination when WEE1 is inhibited. In CDK1- respectively. As expected, Cyclin E led to increased chromatin depleted cells, MK-1775–induced PCNA monoubiquitina- binding of CDC45 (a distal and rate-limiting step in the ini- tion was attenuated when compared with CDK1-replete cells tiation of DNA synthesis) coincident with increased PCNA (Fig. S1 A), possibly suggesting that CDK1 contributes to monoubiquitination (Fig. 2 A). Overexpressed licensing factors PCNA monoubiquitination. However, CDK1-depleted cells did not promote chromatin binding of CDC45 and induced neg- were arrested in G2/M (Fig. S1 B), potentially causing at- ligible changes in PCNA monoubiquitination (Fig. 2 A). There- tenuation of PCNA monoubiquitination indirectly after cell fore, RAD18 activation is associated with aberrant initiation of cycle arrest. Therefore, as an alternative approach to de- DNA synthesis, not altered replication licensing. termining the potential roles of CDK2 and CDK1 in stim- In a complementary approach to test the relationship ulating PCNA monoubiquitination, we tested the effects of between PCNA monoubiquitination and initiation of DNA specific CDK2 activators (Cyclins E and A) or CDK1 acti- synthesis, we used p53-directed shRNA or expressed the high- vators (Cyclin B and CDC25C) on PCNA. As shown in Fig. risk HPV-E6 oncoprotein to abrogate the G1 checkpoint and S1 (C and D), PCNA monoubiquitination was induced by increase initiation of DNA synthesis. As expected, p53 deple- Cyclin E and more modestly by Cyclin A, but not by Cy- tion promoted chromatin binding of CDC45 (Fig. 2 B) and S clin B or CDC25C. We infer that MK-1775–induced PCNA phase entry (Fig. 2 C) while also coincidently inducing PCNA monoubiquitination is primarily caused by de-repression of monoubiquitination (Fig. 2 B). Similarly, acute Wee1 inhibition CDK2 and is not CDK1 mediated. Collectively, Fig. 1 (A–E) in asynchronous HPV-E6–expressing 3T3 cells led to a 4.7-fold and Fig. S1 show that deregulated CDK2 signaling triggers increase in PCNA monoubiquitination when compared with PCNA monoubiquitination in untransformed human cells. p53-replete 3T3 cells (Fig. 2 D). The results of Fig. 2 (B–D) PCNA was also monoubiquitinated in response to aberrant suggest that loss of the G1 checkpoint and the ensuing increases CDK2 activity in other cell types that have been used ex- in S phase entry and origin firing lead to increased PCNA tensively to model cellular responses to oncogenic stress, monoubiquitination. We considered the possibility that p53 loss

for protein concentration and then analyzed by SDS-PAGE and immunoblotting using antibodies against Cyclin E and GAPDH (for loading controls). (C) IP–kinase assay showing relative CDK2 activities in RAS-transformed cells (BJ-RAS) and isogenic parental fibroblasts (BJ). Also shown are CDK2 activities in control, MK-1775–treated, and Cyclin E–overexpressing NHFs and in H1299 lung adenocarcinoma cells. (D) Immunoblot showing the effects of 24 h roscovitine treatment on PCNA monoubiquitination in BJ and BJ-RAS cells. (E) Control and RAD18-depleted NHF cells were treated for 0–7 h with MK-1775. At different times, cell lysates were analyzed for expression of the indicated DNA damage markers using SDS-PAGE and immunoblotting. (F) Rad18+/+ and Rad18−/− MEFs were treated with MK-1775 for 7 h. Lysates from the resulting cells were analyzed by SDS-PAGE and immunoblotted with indicated antibodies. (G) U2OS cells were coinfected with AdCFP-RAD18 and AdCyclin E (or with control adenovirus) for 24 h. Some cultures were UV irradiated (20 J/m2), and 2 h later, CFP-RAD18 subcellular distribution was analyzed by confocal microscopy. Bar, 10 µm. (H) NHFs were transfected with siRAD18 or with nontargeting control siRNA. Transfected cells were infected with adenoviruses encoding siRNA-resistant RAD18, Cyclin E, or with an “empty” virus for control. Lysates from the resulting cells were analyzed by SDS-PAGE and immunoblotted with the indicated antibodies.

3100 JCB • Volume 216 • Number 10 • 2017 Figure 2. Passage through the G1/S restriction point and initiation of DNA synthesis are necessary for CDK2-induced PCNA monoubiquitination. (A) NHFs were infected with the indicated adenoviral vectors for 24 h, and cell lysates were analyzed by SDS-PAGE and immunoblotted with the indicated antibodies. (B) Replicate cultures of NHFs and shp53-NHF cells were transfected with siRAD18 or control nontargeting RNA. The resulting cultures were infected

RAD18 confers oncogenic stress tolerance • Yang et al. 3101 might directly promote PCNA monoubiquitination. However, PCNA monoubiquitination induced by elevated CDK2 activity when we inhibited Wee1 in S phase–synchronized cultures, does not result from limited availability of nucleotide pools. HPV E6 expression did not affect PCNA monoubiquitination CDK2-induced aberrant DNA replication structures are (Fig. 2 E). Moreover, Wee1 inhibition before S phase did not substrates for nucleolytic cleavage by MUS81 (Neelsen et al., induce PCNA monoubiquitination (Fig. 2 E). Therefore, p53 2013a). Beck et al. (2012) showed that MUS81 is recruited to limits CDK2-induced PCNA monoubiquitination by enforcing stalled forks when WEE1 kinase is inhibited. Moreover, MUS81 a G1/S checkpoint and preventing S phase entry. is known to process stalled replication forks to generate the To test whether Cyclin E–induced PCNA monoubiquiti- RPA-ssDNA that triggers CHK1 pathway activation (Regairaz nation was dependent on initiation of DNA replication, we et al., 2011). Therefore, we used gene silencing to test a possi- used pharmacological inhibition of CDK2 and gene silencing ble role for MUS81 in RAD18 pathway activation. As expected of CDC7 to inhibit protein kinase activities that are essential from a previous study (Regairaz et al., 2011), levels of chroma- for origin firing. Fig. 2 (F and G) show that CDK2 and CDC7 tin-bound pRPA (pS4/8) and pChk1 (S317) were reduced after are necessary for Cyclin E–induced PCNA monoubiquitina- MUS81 depletion (Fig. 3 F). Interestingly, MK-1775–induced tion. As expected, silencing of CDC6, which limits the number PCNA monoubiquitination was attenuated in MUS81-depleted of licensed origins available for initiation of DNA synthesis, cells (Fig. 3 F). BrdU labeling rates and cell cycle profiles were also abrogated Cyclin E–induced PCNA monoubiquitination unaffected by MUS81 depletion (Fig. 3 F). Therefore, the re- (Fig. 2 G). We conclude that CDK2-induced PCNA monoubiq- duced PCNA monoubiquitination of MUS81-depleted cells was uitination requires passage through the G1/S restriction point not secondary to an altered S phase. and initiation of DNA synthesis. We used a flow cytometry–based assay to determine relative levels of MK-1775–induced ssDNA in control and MUS81 and replication protein A (RPA) MUS81-depleted cells. The flow-based ssDNA assay is a contribute to RAD18 pathway activation modified version of a previously published protocol for visu- The mechanism of RAD18 activation during S phase has been alizing ssDNA regions (Raderschall et al., 1999) and was vali- attributed to its direct association with p95/NBS1 (Yanagihara dated as described in Fig. S2. As shown in Fig. 3 G, MK-1775 et al., 2011), with RPA-coated single-stranded DNA (ssDNA; led to an approximately twofold increase in ssDNA levels in Davies et al., 2008; Tsuji et al., 2008), or with ubiquitinated MUS81-replete cells. However, WEE1 inhibition did not in- chromatin generated by the E3 ligase RNF8 (Huang et al., duce ssDNA in MUS81-depleted cultures (Fig. 3 G). We con- 2009) at damaged replication forks. clude that certain CDK2-induced aberrant DNA replication Levels of chromatin-bound p95/NBS1 and RNF8 were intermediates can be processed by MUS81 to generate RPA- not increased upon Cyclin E overexpression (Fig. 3 A). More- ssDNA, which in turn serves as a common trigger for both over, depletion of p95/NBS1 or RNF8 did not affect Cyclin E– RAD18 and CHK1 pathways. induced PCNA monoubiquitination (Fig. 3 A), suggesting that In Cyclin E–expressing cells (in which CDK2 activity was p95/NBS1 and RNF8 do not mediate RAD18 pathway activa- ∼2.6-fold higher than in MK-1775–treated cells under our stan- tion in response to excess CDK2. However, Cyclin E overex- dard experimental conditions; see Fig. 1 C), PCNA monoubiq- pression and MK-1775 treatment both led to increased numbers uitination was very modestly reduced after MUS81 depletion of nuclear RPA foci (Fig. 3 B) and increases in chromatin-bound (Fig. 3 H). As noted previously (Neelsen et al., 2013a), the stan- RPA (Fig. 3, A and C), coincident with PCNA monoubiquitina- dard conditions used to express Cyclin E generate severe DNA tion. Moreover, Cyclin E–induced PCNA monoubiquitination replication defects, irreversible cell cycle arrest, and massive was attenuated in RPA-depleted cells (Fig. 3 A). WEE1 inhi- amounts of ssDNA in all cells. Therefore, MUS81-independent bition was previously shown to cause an increase in levels of processing of DNA replication intermediates also contributes ssDNA (Beck et al., 2010) and in alkaline COMET assays, Cy- to PCNA monoubiquitination after the severe replication stress clin E overexpression led to a 27-fold increase in ssDNA levels from overexpressed Cyclin E. relative to control cultures (Fig. 3 D). Collectively, our results are most consistent with an RPA-ssDNA–mediated RAD18 ac- Cyclin E–induced PCNA monoubiquitination tivation mechanism in response to excess CDK2 activity. specifically regulates Polκ Nucleoside depletion has been postulated as a mechanism To identify relevant effectors of genome maintenance in re- for replication defects in response to certain oncogenic stimuli sponse to excess CDK2 signaling, we investigated the regu- (Bester et al., 2011). However, exogenously added nucleosides lation of Y family DNA polymerases. Of the Y family DNA did not suppress the Cyclin E–induced PCNA monoubiquiti- polymerases examined (Polη, Polκ, and Polι), only Polκ was nation (Fig. 3 E). Therefore, we consider it most likely that the redistributed to chromatin in response to overexpressed Cyclin

with the indicated viruses. One plate of each replicate culture was analyzed by SDS-PAGE and immunoblotted with the indicated antibodies. The other plate of each replicate pair was pulse labeled with BrdU for 1 h. The resulting cultures were costained with propidium iodide and BrdU and then analyzed for cell cycle distribution as shown in C. The white dotted line serves to separate the NHF and shp53-NHF data for easier comparison of protein levels between the two cell lines. (C) Graphs summarizing the distribution of cells between different cell cycle phases for each experimental condition. (D) Swiss 3T3-XSN and 3T3-E6 cells were transfected with siRAD18 or with nontargeting siRNA and then treated with the indicated concentrations of MK-1775 for 5 h. Cell lysates were analyzed by SDS-PAGE and immunoblotting with the indicated antibodies. (E) Quiescent (G0) cultures of Swiss 3T3-XSN and 3T3-E6 cells were stimulated with 10% serum. At 8 and 16 h after serum stimulation (time points corresponding to mid-G1 and S phase, respectively), cells were treated with MK-1775 for 5 h before harvest, SDS-PAGE, and immunoblot analysis of the indicated proteins. (F) NHFs were infected with AdCyclin E (or empty vector for control) for 24 h. The resulting cultures were treated with roscovitine for 2 h before harvest for SDS-PAGE and immunoblot analysis. (G) NHFs were transfected with siCDC6, siCDC7, or nontargeting RNA duplexes. The resulting cultures were infected with AdCyclin E (or empty adenoviral vector). 48 h later, cells were harvested for SDS-PAGE and immunoblotting with the indicated antibodies.

3102 JCB • Volume 216 • Number 10 • 2017 Figure 3. CDK2-induced PCNA monoubiquitination is p95/NBS1-independent but requires MUS81. (A) NHFs were transfected with siNBS1, siRPA32, siRNF8, siCHK1, or control nontargeting siRNA. The resulting cultures were infected with AdCyclin E or control adenovirus, and 48 h later, cells were harvested for SDS-PAGE and immunoblot analysis. (B) Replicate cultures of NHFs were infected with AdCyclin E for 24 h, treated with MK-1775 for 5 h, or left

RAD18 confers oncogenic stress tolerance • Yang et al. 3103 E (Fig. 4 A). Cyclin E overexpression also induced association sustain DNA replication and prevent DSB accumulation in Cy- of Polκ with PCNA as determined by coimmunoprecipitation clin E–overexpressing cells. experiments (Fig. 4 B). In immunofluorescence microscopy TLS operates postreplicatively and patches 5′ ssDNA studies, Cyclin E overexpression and WEE1 inhibition both gaps generated behind nascent leading strands (Lopes et al., led to an increase in GFP-Polκ foci (Fig. 4 C). As expected 2006; Daigaku et al., 2010). We predicted therefore that TLS from previous work, YFP-Polη formed constitutive nuclear deficiency would lead to persistent ssDNA when there is ex- foci in replicating cells, but numbers of YFP-Polη nuclear foci cess CDK2. To test this prediction, we used flow cytometry to were unaffected by Cyclin E expression or WEE1 inhibition compare ssDNA levels in Polk+/+ and Polk−/− cells basally and (Fig. 4 D; Durando et al., 2013). Collectively, the results of after Wee1 inhibition. As shown in Fig. 5 E, basal ssDNA levels Fig. 4 suggest that Polκ is recruited to CDK2-induced aberrant were indistinguishable between exponentially growing Polk+/+ DNA replication intermediates. and Polk−/− cells. However, after Wee1 inhibition, Polk−/− MEFs accumulated up to 10-fold higher levels of ssDNA than isogenic Rad18 and Polκ sustain ongoing DNA Polk+/+ cells (Fig. 5 E). The ssDNA induced by Wee1 inhibition replication and cell viability after Wee1 in Polk−/− cells coincided with 4N (G2 + M) populations iden- inhibition tified by propidium iodide staining (Fig. 5 E). Using anti–phos- The results of Figs. 1, 2, 3, and 4 suggested a possible role for pho–histone H3 (to more precisely define mitotic populations), TLS in tolerating the aberrant S phase induced by excessive we demonstrated that mitotic Polk-deficient MEFs accumulated CDK2 activity. To elucidate the contribution of TLS to ongoing approximately fivefold higher levels of ssDNA after Wee1 inhi- DNA synthesis in cells harboring excess CDK2, we compared bition when compared with Polk+/+ cells (Fig. 5 F). We conclude DNA replication dynamics of MK-1775–treated Polk−/− MEFs that Polκ is important to prevent mitotic ssDNA accumulation with isogenic (Polk+/+) control cells. in response to Wee1 inhibition. As shown by the DNA fiber analyses in Fig. 5 (A and B), Persistent ssDNA triggers the “replication checkpoint,” a DNA replication fork velocities were reduced by 15% after 5 h CHK1-mediated mechanism that inhibits CDC2/CDK1 via Y15 of Wee1 inhibition in the Polk+/+ MEFs. However, in Polk−/− phosphorylation and prevents mitosis when the genome is in- cells, Wee1 inhibition led to a 44% reduction of fork velocity. completely replicated (Canman, 2001). In NHFs, RAD18 deple- New origin firing was increased approximately twofold after tion led to an increase in Y15-phosphorylated CDC2 (Fig. 6 A), WEE1 inhibition in both Polk+/+ and Polk−/− MEFs. In NHFs, indicating that the replication checkpoint is activated when TLS increased Cyclin E expression (driven by a stably integrated is compromised. As expected, WEE1 inhibition eliminated the doxycycline-inducible expression vector) led to a 23% decrease Y15 phosphorylation in RAD18-depleted cells (Fig. 6 A). in replication fork velocities (Fig. 5 C). However, in Polκ- We hypothesized that the persistent ssDNA induced by depleted fibroblasts, replication fork velocities were reduced MK-1775 in TLS-deficient cells resulted from aberrant activa- by 63% after Cyclin E induction (Fig. 5 C). Replication fork tion of CDK2 (which deregulates DNA synthesis) and CDK1 velocities were unaffected by Polη depletion (Fig. 5 C). We con- (which abrogates the G2/M checkpoint). To test whether the clude that Polκ-dependent TLS sustains DNA replication fork G2/M checkpoint is a failsafe when TLS is compromised, progression in cells harboring elevated CDK activity. we used ectopically overexpressed CDC25C (a Y15-directed In genotoxin-treated cells, TLS prevents DNA replica- CDK1 phosphatase) as an MK-1775–independent strategy tion fork collapse and DNA DSB accumulation (Bi et al., 2005, to bypass the G2/M checkpoint. As expected, overexpressed 2006). To determine whether TLS also protects against DSB CDC25C caused a reduction in levels of Y15-phosphorylated formation caused by excessive CDK2, we enumerated 53BP1 CDK1 (Fig. S3 A). We determined the effects of CDC25C on nuclear foci (as a surrogate for unrepaired DSBs) in Cyclin mitotic ssDNA in RAS-transformed BJ cells (which have high E–overexpressing cells that were depleted of RAD18, Polκ, or CDK2 activity as shown in Fig. 1 C) and parental BJ fibroblasts. Polη. As shown in Fig. 5 D, silencing of RAD18 or Polκ (but As shown in Fig. S3 C, CDC25C overexpression led to an ∼20- not Polη) led to an approximately twofold increase in the num- fold increase in mitotic ssDNA levels in RAS-transformed cells ber of 53BP1 foci in Cyclin E–overexpressing cells. The results but not in parental BJ fibroblasts. Interestingly, in RAD18- of Fig. 5 (A–D) suggest that RAD18 and Polκ (but not Polη) depleted BJ-RAS cells, levels of mitotic ssDNA were further increased to ∼29-fold above control (Fig. S3 C). In three separate

untreated for controls. The resulting cells were fixed, stained with anti-RPA 32 and DAPI, and analyzed by confocal microscopy. The images represent representative fields of RPA- and DAPI-stained nuclei. At least 100 cells were scored for each experimental condition. Bar, 10 µm. (C) NHFs were transfected with siRAD18 or nontargeting siRNA and then treated with MK-1775 for 5 h. Cells were harvested and analyzed by SDS-PAGE and immunoblotting with the indicated antibodies. (D) NHFs were infected with the indicated adenoviruses for 48 h. Nuclei from the resulting cultures were analyzed using alkaline and neutral comet assays. 50 tail moments were measured for each experimental condition. To determine the statistical significance of the differences in single-strand break (SSB) and DSB levels, we performed ANOV A between groups followed by Tukey’s multiple comparison of means test. Results of the Tukey test indicated significant differences between AdCon and AdCyclin E–overexpressing cells (P < 0.001), indicating that Cyclin E expression induces single-strand breaks and DSBs. The result shown is representative of two independent experiments that yielded qualitatively similar results. ***, P < 0.001. (E) NHFs were grown for 48 h in complete medium or medium supplemented with 50 µM of all four nucleosides (A, U, C, and G) and increasing doses of AdCyclin E (ranging from 0.05−1.0 × 1010 pfu/ml). Chromatin fractions from the resulting cells were analyzed by immunoblotting with the indicated antibodies. (F) NHFs were transfected with siMUS81 and nontargeting control siRNAs for 48 h. The resulting cultures were treated with the indicated doses of MK-1775 for 5 h before harvest for SDS-PAGE and immunoblot analysis with the indicated antibodies. For siMUS81 and siCon-transfected cultures, one replicate plate was pulse labeled with 10 µM BrdU and analyzed by flow cytometry to quantify S phase–positive populations. (G) Replicate plates of NHFs were transfected with the indicated siRNAs for 48 h. One plate of each replicate pair was treated with MK-1775 for 5 h before flow cytometry analysis of ssDNA. (H) U2OS cells were transfected with siMUS81 or nontargeting control (siCon) RNAs. 24 h after transfection, cultures were infected with AdCyclin E or AdCon. After 24 h, cells were collected and analyzed by SDS-PAGE and immunoblotting with the indicated antibodies.

3104 JCB • Volume 216 • Number 10 • 2017 Figure 4. Cyclin E overexpression and WEE1 inhibition promote nuclear redistribution and PCNA binding of Polκ. (A) U2OS cells were infected with 0, 108, and 1010 pfu/ml of AdCyclin E. 48 h later, chromatin extracts from the resulting cells were analyzed by SDS-PAGE and immunoblotting with the indicated antibodies. (B) U2OS cells were infected with AdCyclin E, AdHA-PCNA, and AdGFP-Polκ as indicated. 24 h later, solubilized chromatin fractions were prepared and analyzed directly by SDS-PAGE (“input”) or were first immunoprecipitated with anti-HA (IP) to recover PCNA complexes, which were analyzed for the presence of associate GFP-Polκ. (C and D) Replicate cultures of U2OS cells were infected with GFP-Polκ (C) or YFP-Polη (D) adenovirus for 24 h. The resulting cultures were infected with AdCyclin E for 24 h, were treated with MK-1775 for 5 h, or were left untreated for controls. The subcellular distributions of GFP-Polκ and YFP-Polη were analyzed by fluor escence microscopy, and representative images of TLS polymerase–expressing nuclei are shown. Bars, 10 µm. For quantitative analysis of DNA polymerase foci, at least 100 nuclei were scored for each experimental condition. The results are compiled from three independent experiments. Error bars represent SEM. Statistical analyses were performed using Student’s t test. The differences in numbers of foci between experimental conditions were significant for GFP-Polκ (**, P < 0.01) but not significant for YFP-Polη (P > 0.05).

experiments, RAD18 depletion increased the levels of RAS- when compared with isogenic WT control cells (Fig. 6 C). In induced ssDNA by 1.4-fold ± 0.16 (P = 0.028). Collectively, the clonogenic survival assays, both Rad18−/− and Polk−/− MEFs results of Fig. S3 indicate that in the absence of TLS, CDK2- were sensitive to Wee1 inhibition when compared with WT induced ssDNA persists into mitosis when the replication cell lines (Fig. 6, D and E). We conclude that Rad18 and Polκ checkpoint is bypassed by CDK1. are important for preventing PCC and maintaining viability Forced bypass of the replication checkpoint can elicit when Wee1 is inhibited. premature chromatin condensation (PCC), abortive mitoses To test whether Polκ is specifically required to tolerate (termed “mitotic catastrophe”), and cell death (Nghiem et al., aberrant CDK activity, we also determined the MK-1775 sen- 2001). To determine whether TLS averts mitotic catastrophe, sitivity of an XP-variant (XPV) cell line (XP30RO), which we analyzed metaphase spreads from Rad18−/− and Polk−/− harbors an inactivating mutation in Polη (Volpe and Cleaver, MEFs and isogenic WT cells after MK-1775 treatment and 1995). In contrast with Polk-null MEFs, Polη-deficient XPV scored metaphase chromosomes for PCC. fibroblasts were not MK-1775 sensitive when compared with Images of representative metaphase spreads illustrating isogenic Polη-complemented (XP30RO + Polη) cells. Surpris- normal mitotic chromosomes and pulverized chromosomes ingly, Polη-deficient XP30RO cells were significantly more tol- caused by PCC are shown in Fig. 6 B. In control (MK-1775 erant of WEE1 inhibition when compared with the corrected untreated) MEFs, the incidence of spontaneous PCC was neg- XP30RO + Polη line (Fig. 6 E). To further test the role of Polη ligible (<1%) in WT, Rad18−/−, and Polk−/− cultures, indicating on MK-1775 tolerance, we compared the MK-1775 sensitiv- minimal mitotic defects when TLS alone is compromised. After ity of MEFs derived from transgenic knock-in mice express- MK-1775 treatment, the incidence of PCC increased to 14% ing WT RAD18 or a Polη interaction-deficient RAD18 DC2 and 8% in Rad18+/+ and Polk+/+ MEFs, respectively (Fig. 6 C). mutant (Yang et al., 2016). Cells expressing RAD18 DC2 are Remarkably, the numbers of nuclei undergoing PCC were 3.1- compromised for Polη activation and have modest UV sensi- and 4.4-fold higher in Rad18−/− and Polk−/− MEFs, respectively, tivity (Watanabe et al., 2004; Day et al., 2010). Interestingly,

RAD18 confers oncogenic stress tolerance • Yang et al. 3105 Figure 5. Polκ facilitates DNA replication and prevents ssDNA accumulation in cells with elevated CDK2 activity. (A and B) DNA fiber analyses showing effects of MK-1775 treatment on DNA replication dynamics in cultures of isogenic WT and Polk−/− MEFs. Panel A shows results of a representative experiment in which fork velocities were determined for 50 individual replication tracts under each experimental condition. Based on the results of ANOVA and

3106 JCB • Volume 216 • Number 10 • 2017 similar to Polη-deficient XPV cells, RAD18 DC2 MEFs were ance of CDK2-induced DSBs. However, inactivating the Polq MK-1775 resistant when compared with WT RAD18-express- gene did not increase the MK-1775 sensitivity of Rad18−/− or ing MEFs (Fig. 6 G). Therefore, Polκ is specifically required to Polk−/− MEFs (Fig. S4). Collectively, these results indicate tolerate excessive CDK activity, and Polη inhibits tolerance of that there is partial redundancy between TLS and HR (but CDK2-induced replication stress. not between TLS and TMEJ) for MK-1775 tolerance. These To extend our analysis of Rad18 functions in tolerance findings validate RAD18-Polκ as a therapeutic target path- of excess CDK2 activity, we also determined the role of Rad18 way for augmenting MK-1775–induced lethality in cancer in cellular response to oncogenic RAS. Rad18+/+ and Rad18−/− cells, particularly those with HR defects. MEFs were transduced with retroviruses encoding oncogenic RAS or with an empty vector. As shown in Fig. 6 H, growth curves of empty vector–transduced Rad18+/+ and Rad18−/− cells Discussion were identical. However, RAS expression stimulated proliferation of Rad18+/+ MEFs but was inhibitory for Rad18−/− cells. It is widely accepted that diverse oncogenes induce DNA These results further demonstrate that Rad18 facilitates toler- replicative stresses that in turn activate canonical DNA dam- ance of oncogenic stress. age signaling and repair pathways (Halazonetis et al., 2008). Many bona fide oncogenes including RAS induce CDK2 ac- Rad18 deficiency sensitizes cancer cells tivity (Fikaris et al., 2006). The high-level CDK2 activity of and Brca1-deficient cells to Wee1 inhibition oncogene-expressing cells is a likely source of DNA replication Because MK-1775 is being considered as a therapeutic agent, stress. We show in this study that three independent CDK2-acti- we determined the extent to which RAD18 signaling allows vating stimuli (oncogenic RAS, Cyclin E, and WEE1 inhibition) cancer cell lines to tolerate WEE1 inhibition. As shown in induce a DNA damage response that involves TLS pathway Fig. 7 (A–C), three representative and commonly studied can- activation. Interestingly, Soria et al. (2006) previously showed cer cell lines, H1299, A549, and U2OS, were MK-1775 sensi- that p21–CDK2 interactions suppress PCNA monoubiquitina- tive after RAD18 depletion. tion and speculated that CDK activity may promote TLS—a Sensitivity to replicative stress is likely determined by prediction borne out by our results. the repertoire of available DNA damage tolerance pathways. The precise nature of the aberrant DNA replication struc- Therefore, we sought to identify the DNA repair mechanisms tures induced by oncogenic stimuli or excess CDK activity is that cooperate with RAD18 to confer tolerance of MK-1775– not well understood. Rereplication resulting in initiation of induced stress. TLS deficiency predisposes to DSB forma- DNA synthesis more than once per S phase may be a major tion (Fig. 5 D). Therefore, we determined the contributions mechanism of oncogene-induced DNA replication stress and of the two major S phase–coupled DSB repair processes (ho- DNA damage (Bartkova et al., 2006; Di Micco et al., 2006). mologous recombination [HR] and θ-mediated end joining Rereplication induces ssDNA gaps and replication fork reversal [TMEJ]) to MK-1775 tolerance in Polk-deficient cells. We (Neelsen et al., 2013b) that eventually lead to DSB and check- used lentivirally expressed single guide RNAs (sgRNAs) to point activation (Vaziri et al., 2003). The ssDNA generated by functionally inactivate Brca1 or Polq genes (encoding essen- rereplicating DNA provides a potential mechanism for acti- tial mediators of HR and TMEJ, respectively) in Polk+/+ or vation of both CHK1 and RAD18 pathways. However, CDT1 Polk−/− MEFs. As shown in Fig. 7 D, disruption of Brca1 overexpression or geminin depletion—treatments that induce led to increased MK-1775 sensitivity of both Polk+/+ and massive relicensing and rereplication (Vaziri et al., 2003)— Polk−/− MEFs, indicating that the HR and TLS jointly (and triggered modest PCNA ubiquitination when compared with independently) facilitate tolerance MK-1775–induced DNA Cyclin E overexpression or MK-1775 treatment. Therefore, damage. As shown in Fig. S4, Polq-deficient MEFs were rereplication intermediates may not be the sole mediators of MK-1775 sensitive, also indicating a role for TMEJ in toler- CDK2-induced TLS pathway activation.

Tukey’s multiple comparison of means test, MK-1775 treatment led to statistically significant decreases in fork velocity inPolk −/− MEFs (P < 0.001) but not Polk+/+ cells (P > 0.05). To quantify new origin firing, 300 fibers were counted for each experimental condition. There were no significant differences in numbers of new origins when comparing between Polk−/− and Polk+/+ MEFs. In panel B, lengths of ongoing replication forks (containing both CldU and IdU) were measured to determine the distribution of replication fork speeds. 50 fibers were quantified for each experimental condition. The experiment described in A and B was repeated twice with similar results. (C) DNA fiber analyses showing effects of doxycycline (Dox)-inducible Cyclin E expression on DNA replication dynamics of Polκ- or Polη-depleted NHFs. DNA replication fork velocities were determined for 150 individual replication tracts from each experimental condition. Based on the results of ANOV A and Tukey’s multiple comparison of means test, doxycycline-inducible Cyclin E expression led to statistically significant reduction of fork velocity by in siPOLK-transfected NHFs (P < 0.001) but not siPOLH-transfected cells (P > 0.05). The data presented are representative of an experiment that was repeated twice with similar results. (D) Replicate cultures of U2OS cells were transfected with the indicated siRNAs. After 48 h, cells were treated with 2.5 µM MK-1775 (or vehicle for control) for 5 h. Cells were fixed, and 53BP distribution was analyzed by immunofluorescence confocal microscopy. Representative images of 53BP1-stained nuclei are shown. Bar, 10 µm. For quantitative analysis of DNA, at least 100 nuclei were scored for each experimental condition. The graph shows results of three independent experiments, and error bars represent SEM. Statistical analyses were performed using a Student’s t test. The resulting p-values indicate significant differences in numbers of 53BP1 foci between siCon and siPOLK (P < 0.05) and between siCon and siRAD18 (P < 0.01), but no significant differences between siCon and siPOLH-transfected cells. (E) Replicate cultures of isogenic Polk+/+ and Polk−/− MEFs were treated with 2.5 µM MK-1775 for the indicated times and then analyzed for ssDNA content using flow cytometry. For the 0- and 5-h time points, one plate of each replicate culture was collected for SDS-PAGE and immunoblotting with the indicated antibodies. PI, propidium iodide. (F) Polk+/+ and Polk−/− MEFs were prelabeled with BrdU for 24 h and then treated with 2.5 µM MK-1775 (or vehicle) for 5 h. Cultures were analyzed by flow cytometry to identify phospho–histone H3– and ssDNA-positive populations. The histogram shows the results from three independent experiments in which the columns represent means and error bars represent the range. From the results of ANOVA and Tukey’s multiple comparison of means test, MK-1775 treatment induced statistically significant increases in ssDNA levels inPolk −/− MEFs (P < 0.001) but not in Polk+/+ cells (P = 0.8). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

RAD18 confers oncogenic stress tolerance • Yang et al. 3107 Figure 6. Polκ is specifically required to prevent PCC and confer viability after WEE1 inhibition.(A) U2OS cells were transfected with siRAD18, siWEE1, or nontargeting control RNA. 48 h later, cells were analyzed by SDS-PAGE and immunoblotting with the indicated antibodies. (B and C) Replicate cultures of isogenic Polk+/+ and Polk−/− MEFs were treated with 2.5 µM MK-1775 (or vehicle) for 7 h. Metaphase spreads from the resulting cells were analyzed by light microscopy and scored for aberrant nuclei undergoing mitotic catastrophe. Molecular masses are given in kilodaltons. Panel B shows representative examples of normal metaphase spreads and of pulverized nuclei displaying hallmarks of PCC. Bars, 10 µm. The histogram in C shows the effect of Polk and Rad18 on the relative incidence of PCC in control and MK-1775–treated cells. The results shown in the histogram are derived from three independent experiments in which at least 200 nuclei were scored for every experimental condition. The error bars represent SEM. Statistical analyses using ANOV A and Tukey’s multiple comparison of means test show significant differences in the percentage of nuclei undergoing PCC when comparingRad18 +/+ versus Rad18−/− MEFs (P < 0.01) and also when comparing Polk+/+ and Polk−/− cells (P < 0.01). There was no significant difference betweenRad18 −/− and Polk−/− cells basally or after MK-1775 treatment (P > 0.05), indicating that Polk−/− cells recapitulate the PCC phenotype of Rad18-null MEFs. (D–G) The indicated pairs of isogenic cell lines with mutations in Rad18, Polk, and POLH were treated with different doses of MK-1775, and sensitivity to Wee1/WEE1 inhibition was evaluated by clonogenic survival assays. For each experiment, the number of surviving colonies from MK-1775–treated cultures was expressed as a percentage of colony numbers from cells that received vehicle (DMSO) for control. On the survival curves, each data point represents the mean of triplicate determinations, and error bars represent the range. For each experiment, we performed a Student’s t test with the data obtained using 0.25 µM MK-1775. When comparing Rad18+/+ versus Rad18−/− cells or Polk+/+ versus Polk−/− cells, p was <0.01 (indicating that Rad18−/− and Polk−/− MEFs are MK-1775 sensitive when compared with their respective WT control cell lines). When comparing XP30RO and XP30RO + Polη cells, p was <0.05, indicating that Polη-complemented XPV cells are MK-1775 sensitive when compared with Polη-deficient XPV patient cells. When comparing “knock-in” MEFs expressing RAD18 WT and RAD18 DC2, p was <0.01, indicating that MEFs expressing the RAD18 DC2 mutant cells are more MK-1775 tolerant than WT RAD18- expressing cells. All data shown are from representative experiments that were repeated at least three times and yielded similar results on each occasion. (H) Primary untransformed Rad18+/+ or Rad18−/− MEFs were cultured in 3% oxygen and infected with pMXs-derived retrovirus encoding HRASV12 or with empty retroviral vector for control. Retrovirus-infected cells were seeded in medium containing puromycin (1 µg/ml), and puromycin-resistant cells were enumerated every day for 1 wk. Data points represent means of triplicate determinations with error bars representing SEM. Using the data from the day 4 post-infection time point, we performed ANOV A between groups followed by Tukey’s multiple comparison of means test to evaluate statistical significance of differences in growth of various genotypes. Results of the Tukey test were as follows: Rad18+/+ (empty vector) versus Rad18−/− (empty vector), P > 0.05 (indicating no significant difference in proliferation ofRad18 +/+ and Rad18−/− in the absence of oncogenic RAS); Rad18+/+ (empty vector) versus Rad18+/+ (RASV12), P < 0.01 (indicating significant stimulation of proliferation by oncogenic RAS inRad18 -replete cells); Rad18−/− (empty vector) versus Rad18−/− (RASV12), P < 0.01 (indicating significant inhibition of proliferation by RASV12 in Rad18-deficient cells). The cell proliferation data are from a representative experiment that yielded similar results on three separate occasions. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

3108 JCB • Volume 216 • Number 10 • 2017 Figure 7. RAD18 confers MK-1775 tolerance in multiple cancer cells and cooperates with DSB repair genes to promote cell viability. (A–C) The indicated cancer cell lines were transfected with siRAD18 or control nontargeting siRNAs. 48 h later, cells were treated with the indicated doses of MK-1775, and drug sensitivities were determined using clonogenic survival assays. For each experiment, the number of surviving colonies from MK-1775–treated cultures was expressed as a percentage of colony number from cells that received vehicle (DMSO) for control. On the survival curves, each data point represents the mean of triplicate determinations, and error bars represent the range. For each experiment, we performed a Student’s t test with data obtained using the 0.3-µM MK-1775 concentration. When comparing siCon and siRAD18, p was <0.01 for all three cell lines, indicating that these cancer cells depend on RAD18 for MK-1775 tolerance. All data shown are from representative experiments that were repeated at least three times and yielded similar results on each occasion. (D) Cultures of isogenic Polk+/+ and Polk−/− MEFs were infected with lentiviruses encoding nontargeting control guide RNA or sgRNA targeting the murine Brca1 gene. After lentiviral transduction, pools of puromycin-resistant cells were treated with different concentrations of MK-1775, and sensitivity to Wee1 inhibition was evaluated by clonogenic survival assays. For each experiment, the number of surviving colonies from MK-1775–treated cultures was expressed as a percentage of colony number from cells that received vehicle (DMSO) for control. On the survival curves, each data point represents the mean of triplicate determinations, and error bars represent the range. For each dose of MK-1775, we performed ANOVA between groups followed by Tukey’s multiple comparison of means test. From results obtained with 0.2 µM MK-1775, either Polk or Brca1 deficiency led to MK-1775 sensitivity (P = 0.006 for both genotypes when compared with Polk- and Brca1-replete controls), but there was no significant difference in the MK-1775 sensitivity of Polk- and Brca1-deficient cells (P = 0.06). Combined loss ofPolk and Brca1 led to increased MK-1775 sensitivity when compared with individual deficiencies inPolk −/− or sgBrca1 cells (P = 0.001), indicating that Brca1 and Polk are nonepistatic and have redundant roles in MK-1775 tolerance. The graph shown is from a single experiment that was repeated three times with similar results on each occasion. **, P < 0.01.

Regardless of the proximal DNA replication structures cells is caused by relief of competition with Polκ for access to that give rise to TLS pathway activation, ssDNA is clearly replicating DNA. Regardless of whether Polη competes with generated in response to aberrant CDK2 activity (Neelsen et Polκ for CDK2-induced replication intermediates, our results al., 2013a) and most likely initiates TLS pathway activation. suggest a novel tumor-suppressive role for Polη in sensitizing In many of our experiments, CDK2 led to increased chromatin cells to oncogenic DNA replication stress. Conversely, Polκ loading of PCNA (in addition to PCNA monoubiquitination). supports viability in the face of oncogenic stress and may Similarly, both PCNA and TLS polymerases are recruited to facilitate tumorigenesis. oxidative damage-induced ssDNA (Yang et al., 2013). It is es- What, then, is the basis for Polκ dependency of MK-1775 tablished that TLS can remediate post-replicative ssDNA gaps tolerance, and what are the putative DNA substrates that are including ssDNA arising outside of S phase (Daigaku et al., preferentially replicated by Polκ? Persistent ssDNA is known to 2010; Zlatanou et al., 2011; Yang et al., 2013). Genomes con- generate G4 structures (Cea et al., 2015), and TLS polymerases taining persistent ssDNA are vulnerable to nucleolytic attack are implicated in the bypass of quadruplex DNA (Bétous et and are likely to generate lethal DSBs. Collectively, our results al., 2009). Polκ preferentially binds G4 when compared with indicate that RAD18-mediated repair synthesis prevents accu- non-G4 DNA, and Polκ activity is enhanced when the enzyme mulation of CDK2-induced ssDNA gaps in the genome. is within two to three nucleotides of a G4 motif (Eddy et al., Of the Y family DNA polymerases, Polη is the most ver- 2016). However, Polk−/− MEFs did not show increased sensitiv- satile enzyme and is probably the default TLS polymerase re- ity to the G4 stabilizer pyridostatin (PDS) when compared with cruited to most stalled DNA replication forks (Watanabe et al., Polk+/+ cells (Fig. S5). Therefore, reduced tolerance of quadru- 2004). It is very clear that Polη resides constitutively in rep- plex DNA cannot explain the MK-1775 sensitivity of Polk−/− lication factories of unperturbed cells and facilitates tolerance cells. Interestingly, Rad18−/− cells were highly PDS sensitive, of intrinsic replication stresses at fragile sites, telomeres, and indicating that a Rad18 effector other than Polκ is necessary for probably elsewhere in the genome (Bétous et al., 2009; Ber- G4 tolerance. Collectively, our results along with several recent goglio et al., 2013; Garcia-Exposito et al., 2016). It is unex- studies (Despras et al., 2016; Garcia-Exposito et al., 2016) show pected therefore that Polκ but not Polη is specifically required that the different TLS pols have separable roles in processing for tolerance of CDK-induced replication stress. Moreover, it is distinct forms of intrinsically arising DNA replication stress. surprising that Polη inhibits MK-1775 tolerance. When com- Whether TLS polymerases evolved solely to cope with pared with other TLS polymerases, Polη has a high affinity for bulky chemically induced DNA lesions has been a subject of PCNA conferred by a PLTH motif immediately flanking the extensive debate. In this regard, we note that the TLS-defective C-terminal PCNA-interacting peptide box motif (Hishiki et al., (Rad18 and Polk mutant) cell lines used in this study are only 2009; Durando et al., 2013; Despras et al., 2016). We speculate modestly sensitive to UV light, and other bulky lesion-inducing that the increased MK-1775 tolerance of Polη-compromised agents yet show remarkable hypersensitivity to MK-1775. Our

RAD18 confers oncogenic stress tolerance • Yang et al. 3109 Figure 8. Roles of Rad18 and Polκ in tolerance of CDK2-induced DNA replication stress. MK-1775 treatment de-represses CDK2 and CDK1 activities, which induce DNA replication stress and aberrant G2/M progression, respectively. Rad18 and Polκ prevent accumulation of post-replicative ssDNA gaps arising from aberrant CDK2 activity. In the absence of Rad18 or Polκ, CDK2-induced ssDNA persists into mitosis (because of aberrant CDK1 activity) and triggers PCC.

results are consistent with the view that TLS serves primarily to in BRCA1 mutant breast and ovarian cancers (Ceccaldi et al., respond to replicative stresses such as those induced by CDK2. 2015). It will eventually be important to systematically define CDK2-induced oncogenic stress is often interchange- the roles of Rad18, Polk, and Polq in carcinogenesis in response ably modeled in cultured cells using overexpressed Cyclin E or to defined oncogenic drivers using mouse models. MK-1775 treatment (Sogo et al., 2002; Beck et al., 2010, 2012; An important byproduct of the selective pressure for DNA Neelsen et al., 2013a). However, these experimental approaches damage tolerance during tumorigenesis is the emergence of che- differ considerably in that MK-1775 de-represses both CDK2 moresistant cancer cells. If cancer cells are commonly depen- and CDK1 (which promote S phase and G2/M progression, re- dent on pathological DNA repair (via up-regulated TLS, TMEJ, spectively), whereas Cyclin E activates only CDK2. Our stud- or other mechanisms), those genome maintenance mechanisms ies with WEE1 inhibition and CDC25C overexpression reveal represent molecular vulnerabilities that could be harnessed to a requirement for an intact G2/M restriction point to repress sensitize cells to intrinsic or therapy-induced DNA damage. phenotypic defects (such as persistent ssDNA in mitosis and WEE1 inhibition is currently being considered as a therapeu- PCC) that arise if TLS fails. Historically, the UV sensitivities of tic strategy in cancer (Sakurikar et al., 2016). Our data show some XPV cell lines were only evident after caffeine treatment that the anticancer activity of MK-1775 and other WEE1 inhib- (Maher et al., 1976). Based on our results, the requirement for itors could be improved with concurrent inhibition of the TLS caffeine to reveal DNA damage sensitivity is most likely caused pathway. In particular, our results suggest that BRCA1-deficient by inhibition of the ATR–CHK1 pathway and override of the cells will be highly MK-1775 sensitive when TLS is inhibited. G2/M replication checkpoint. Our results reveal extensive interactions between TLS and multiple cell cycle and genome maintenance pathways: We demonstrate that p53 enforces G1 arrest Materials and methods and suppresses TLS activation in cells experiencing CDK2- induced replication stress. The G2/M replication checkpoint Cell culture and transfection represses mitotic entry of TLS-defective cells containing un- hTERT-expressing NHFs were provided by W. Kaufmann (University derreplicated DNA (Fig. 8). Finally, our analysis of interactions of North Carolina at Chapel Hill, Chapel Hill, NC). hTERT-expressing between Rad18/Polk and DSB repair genes (Polq and Brca1) in- NHFs are diploid untransformed cells with no known mutations in on- dicates that there is partial redundancy between TLS and HR for cogenes or tumor suppressors. Primary MEFs were derived from E13.5 tolerance of CDK2 activity. Thus, TLS is a single component of embryos of WT, Rad18−/−, Polk−/−, HA-RAD18-WT knock-in, and HA- a genome maintenance network comprising multiple pathways RAD18 (DC2) knock-in C57/BL6 mice as previously described (Bi et that collectively determine oncogenic stress tolerance. al., 2005; Yang et al., 2016). Polq−/− and isogenic Polq+/+ MEFs were Current paradigms suggest that diverse oncogenic stresses provided by R. Wood (The University of Texas, Smithville, TX). With must be tolerated for neoplastic cells to bypass tumor-suppres- the exception of the engineered mutations in DNA repair genes, the sive barriers (Halazonetis et al., 2008). Our results suggest that MEFs used in this study are not known to harbor additional genetic RAD18/Polκ-mediated TLS could endow neoplastic cells with lesions. Cancer cell lines H1299, A549, U2OS, and 293T were pur- two important tumorigenic phenotypes: DNA damage/replica- chased from ATCC and used for the described experiments without fur- tion stress tolerance and error-prone (mutagenic) DNA synthe- ther authentication. H1299 is a nonsmall cell lung carcinoma–adherent sis. Therefore, tumorigenesis may impose selective pressure epithelial cell line that lacks p53 protein. A549 is a nonsmall cell lung for TLS proficiency. It is reported that Polκ is up-regulated in carcinoma–adherent epithelial cell line which expresses WT p53 pro- tumors (Bavoux et al., 2005), and our previous identification of tein. U2OS is an osteosarcoma cell line with adherent epithelial mor- a cancer cell–specific RAD18 activation mechanism (Gao et al., phology that is intact for p53 and RB signaling. The 293T cell line was 2016a,b) is also consistent with a role for TLS as a carcinogenic derived from the embryonic kidney and was transformed by sequential driver. Other DNA repair pathways are also elevated in some stable expression of the adenovirus E1A and E1B genes and SV40 large cancers. For example, POLQ (which we show in this study T antigen. NHFs harboring a doxycycline-inducible human Cyclin E can confer tolerance of excess CDK2 activity) is up-regulated cDNA were generated using the pINDUCER20 lentiviral vector (Meer-

3110 JCB • Volume 216 • Number 10 • 2017 brey et al., 2011). XP30RO XPV and XP30RO + Polη fibroblast cells IP and immunoblotting were provided by G. Stewart (University of Birmingham, Birmingham, To prepare extracts containing soluble and chromatin-associated pro- England, UK). XP30RO cells are UV sensitive, owing to homozygous teins, monolayers of cultured cells typically in 10-cm plates were deletion of the 5′ region of the POLH gene, leading to expression of washed twice in ice-cold PBS and lysed in 500 µl of ice-cold cyto- truncated and inactive Polη protein (Volpe and Cleaver, 1995). With skeleton buffer (CSK buffer; 10 mM Pipes, pH 6.8, 100 mM NaCl, the exception of mutations in the POLH gene, XP30RO cells are not 300 mM sucrose, 3 mM MgCl2, 1 mM EGTA, 1 mM dithiothreitol, + known to harbor any genetic lesions. The XP30RO Polη cell line is a 0.1 mM ATP, 1 mM Na3VO4, 10 mM NaF, and 0.1% Triton X-100) corrected derivative of XP03O cells that stably expresses WT Polh and freshly supplemented with protease inhibitor cocktail and PhosSTOP displays normal UV sensitivity (Kannouche et al., 2001). All mouse and (Roche). Lysates were centrifuged at 4,000 rpm for 4 min to remove human fibroblasts and cancer cell lines were cultured in DMEM sup- the CSK-insoluble nuclei. The detergent-insoluble nuclear fractions plemented with 10% FBS and 1% penicillin–streptomycin. XP30RO were washed once with 1 ml of CSK buffer and then resuspended in and XP30RO+ Polη cells were cultured in RPMI 1640 medium supple- a minimal volume of CSK before analysis by SDS-PAGE and immu- mented with 10% FBS and 1% penicillin–streptomycin. All cell lines noblotting. For all IP experiments, input samples were sonicated and tested negative for mycoplasma contamination as determined using the normalized for protein concentration. Magnetic beads containing cova- Universal Mycoplasma Detection kit (301012K; ATCC). Plasmid DNA lently conjugated antibodies against the HA tag (Cell Signaling Tech- and siRNA oligonucleotides were transfected using Lipofectamine nology) were added to the extracts, and incubations were performed for 2000 (Invitrogen) according to the manufacturer’s instructions except 30 min at 4°C using rotating racks. Immune complexes were recovered that plasmid DNA and Lipofectamine 2000 concentrations used in each using magnetic stands. The beads were washed five times with 1 ml transfection reaction were decreased by 50% to reduce toxicity. CSK (1 min per wash) to remove nonspecifically associated proteins. The washed immune complexes were boiled in protein loading buffer Adenovirus construction and infection for 10 min to release and denature for SDS-PAGE. The Tet-regulated Cyclin B adenovirus was constructed as previously For immunoblotting, cell extracts or immunoprecipitates were described by Jin et al. (1998). All other adenoviruses were constructed separated by SDS-PAGE, transferred to nitrocellulose membranes, and purified as described previously (Vaziri et al., 2003). In brief, and incubated overnight with the following primary antibodies: PCNA cDNAs encoding various cell cycle regulators (including cyclins, (sc-56), Chk1 (sc-7898), β-actin (sc-130656), cyclin E (sc-198), GAP oncogenes, and DNA replication and repair factors) were subcloned DH (sc-32233), Rnf8 (SC-134492), Cdc6 (SC-9964), and GFP (SC- into the pACCMV shuttle vector. The resulting shuttle vectors were 9996) from Santa Cruz Biotechnology, Inc.; Cdt1 (A300-786A), cotransfected with the pJM17 adenovirus plasmid into 293T cells. Re- pRPA32 S4/8 (A300-245A), Polη (A301-231A), Polι (A301-304A), combinant adenovirus clones were isolated by plaque purification and Polκ (A301-977A), and RAD18 (A301-340A) from Bethyl Labo- verified by restriction analysis and Southern blotting. The empty vector ratories, Inc.; p42 MAPK (9107), p-Cdc2 Y15 (9111), p-Chk1 S317 AdCon (used to control for adenovirus infections) was derived simi- (2344), and p-Chk2 (2661) antibodies were purchased from Cell Sig- larly but by cotransfection of the parental pACCMV shuttle vector with naling Technology; γH2AX (05-636) and RPA34/RPA32 (NA19L) pJM17. Adenovirus particles were purified from 293T cell lysates by were from EMD Millipore; CDC7 (DCS-342) was from MEDICAL polyethylene glycol precipitation, CsCl gradient centrifugation, and gel & BIOLOGICAL LABORATORIES CO.; and Cdc45 rat monoclo- filtration column chromatography. Adenovirus preparations were quan- nal antibody was as previously described (Liu et al., 2006). Antibody tified by 260A measurements. Cells were typically infected with 0.1−1.0 dilutions used for immunoblotting were 1:1,000 with exceptions for × 1010 pfu/ml (as indicated in Figs. 1, 2, 3, and 4) by direct addition of the following antibodies: PCNA (1:500), β-actin (1:5,000), GAPDH purified virus to the culture medium. (1:5,000), and γH2AX (1:10,000).

RNA interference CDK2 activity assays For cancer cell lines, siRNAs were reverse transfected using Lipo- CDK2 activities were determined by IP–kinase assays as described by fectamine 2000. In brief, siRNAs were incubated with Lipofectamine Rosenblatt et al. (1992). In brief, cells were lysed in CSK as described 2000 and serum-free OptiMEM for 15 min at room temperature in the in the IP and immunoblotting section. Lysates were clarified by centrif- dark. Cells were then trypsinized and resuspended in 1 ml of Opti- ugation (10,000 g for 10 min), normalized for protein content (typically MEM and added directly into the siRNA/OptiMEM/Lipofectamine ∼500 µg protein in 0.5 ml), and immunoprecipitated with anti-CDK2 solution to give a plating density of 50%, and then they were incubated antibody (sc-163) or control IgG using magnetic beads as described for 48 h. For NHFs and mouse fibroblasts, siRNAs were transfected in Fig. 1 C. After the final CSK wash, beads were washed in 50 mM using electroporation with a GenePulser Xcell (Bio-Rad Laborato- Hepes, pH 7.4, and 1 mM dithiothreitol and then combined with a 30-µl ries). Electroporation was performed according to the manufacturer’s reaction mix containing 10 µg of histone H1 (Sigma-Aldrich), 50 µM 7 32 instructions. 200 µl PBS containing 10 cells and 5 µM siRNA was ATP, 10 mM MgCl2, and 2.5 µCi of [γ- P]ATP (3,000 Ci/mmol; NEN). electroporated in a 0.2-cm cuvette using a 150 V, 10 ms, and 1 pulse After a 15-min incubation at room temperature, reaction mixtures were for NHFs and using the preset 3T3 program (160 V and 500 µF) for boiled in Laemmli sample buffer and then resolved by SDS-PAGE mouse fibroblasts. Sequences of custom siRNA oligonucleotides used using 12% gels. After electrophoresis, gels were fixed in 40% meth- in this study are as follows: control nontargeting siRNA, 5′-UAGCGA anol/10% acetic acid containing 0.05% Coomassie blue dye and then CUAAACACAUCAA-3′; RAD18, 5′-GAGCAUGGAUUAUCUAUU were washed extensively in 40% methanol/10% acetic acid to remove CAAUU-3′; RNF8, 5′-GAGAAGCUUACAGAUGUUU-3′; RPA32, unincorporated radioisotope. Phosphorylated histone bands were visu- 5′-GGCTCCAACCAACATTGTT-3′; NBS1, 5′-GUACGUUGUUGG alized by autoradiography of the dried gels. AAGGAAA-3′; Chk1, 5′-GCGUGCCGUAGACUGUCCA-3′; Polκ, 5′-GUAAAGAGGUUAAGGAAA-3′; Polη, 5′-GCAGAAAGGCAG COMET assay AAAGUUA-3′; Cdc7, 5′-GCUCAGCAGGAAAGGAGUUdTdT-3′; Relative levels of DNA strand breaks were measured by a single-cell and Cdc6, 5′-ACAAUUAAGUCUCCUAGCA-3′. siCDK1 was the gel electrophoresis assay (Olive et al., 1990) using a commercially ON-TARGETplus SMARTpool siRNA from GE Healthcare. available COMET Assay kit (Trevigen). For COMET assays, NHF

RAD18 confers oncogenic stress tolerance • Yang et al. 3111 cells were plated in six-well dishes at a density of 0.5 × 105 cells per viral transduction, pools of puromycin-resistant cells were seeded in well. Oncogenes (Ha-RAS, MYC, Cyclin E, and CDT1) were delivered six-well plates and treated with different concentrations of MK-1775. to cells by infection with 1010 pfu/ml of purified adenoviral vectors. Cell sensitivity to Wee1 inhibition was evaluated by clonogenic sur- After 24 h, cells were embedded in agarose and subjected to electro- vival assay. For experiments in the cancer cell lines, siRNA transfec- phoresis according to the manufacturer’s instructions. Cells were visu- tions were performed as described in the RNA interference section. alized using an IX61 inverted microscope (Olympus), and images of 24 h after transfection, cells were seeded at a density of 1,000 cells/well cells were acquired as TIFF files. For each experimental condition, “tail in triplicate six-well dishes. MK-1775 treatments and colony-forming moments” (defined as the product of tail length and the fraction of total assays were performed exactly as described above. DNA in the tail) were determined for 50 nuclei using ImageJ software (National Institutes of Health) with the COMET assay plugins (original Flow cytometry macro from Herbert M. Geller, added by Robert Bagnell). To compare To detect ssDNA, growing cells were first cultured for 24 h in medium tail moments between experimental groups, we performed ANOVAs containing 10 µM BrdU to label genomic DNA. Cell monolayers were followed by Tukey’s tests to correct for experiment-wide error rates washed to remove unincorporated BrdU, placed in fresh BrdU-free me- between multiple comparisons. dium, and returned to the incubator. After treatment with genotoxins or oncogenes as described in all figure legends, cells were harvested Analysis of DNA replication dynamics using DNA fiber assays by trypsinization, washed in PBS, resuspended in 65% PBS with 35% Growing cells were pulse labeled for 20 min with 25 µM of the thymi- ethanol, and fixed by overnight incubation at 4°C. Fixed cells were dine analogue chlorodeoxyuridine (CIdU; C6891; Sigma-Aldrich). At stained with fluorescent anti-BrdU antibodies (FITC mouse anti-BrdU the end of the CldU-labeling period, cells were washed twice with a kit; 556028; BD) without prior acid treatment to detect only ssDNA. warm medium and immediately pulse labeled for 20 min with 250 µM To confirm that the experimental cells had incorporated BrdU, aliquots of idodeoxyuridine (IdU; I7125; Sigma-Aldrich). Labeled cells were of all nuclei were also denatured using HCl and then neutralized with harvested and DNA fiber spreads were prepared and stained exactly borax before staining with anti-BrdU antibody. For cell cycle analysis, as described previously by Jones et al. (2013). Images of stained DNA nuclei were incubated in 1× PBS containing 10 µg/ml of propidium io- fibers were acquired on a 710 confocal microscope (UNC Microscopy dide and 1 µg/ml of RNaseA to stain total DNA. Stained cells were an- Services Laboratory; ZEISS) using a 20× lens. The lengths of CldU alyzed by flow cytometry on an Accuri C6 flow cytometer (BD) using (AF 555; red)- and IdU (AF 488; green)-labeled fibers were measured the manufacturer’s software. using the ImageJ software, and micrometer values were converted into kilobases using the conversion factor 1 µM = 2.94 kb (one bp corre- Fluorescence microscopy sponds to ∼340 pm). 50 representative DNA fibers were measured for To visualize nuclear foci containing TLS proteins, U2OS cells were each experimental condition. Prism 6 (GraphPad Software) was used grown to ∼50% confluency on glass-bottomed plates (MatTek Corpora- for variation analysis and to generate dot plots. Fiber distribution fig- tion) then infected with AdYFP-Polη (5 × 108 pfu/ml), AdGFP-Polκ (109 ures were generated using the RColorBrewer package in R. pfu/ml), AdCFP-RAD18 (4 × 109) pfu/ml, or equivalent concentrations of an “empty” control adenovirus (AdCon). To induce replication stress Clonogenic survival assays by Cyclin E overexpression, cells were coinfected with 109 pfu/ml of Ad- For experiments in MEFs, cells were seeded at a density of 1,000 cells/ Cyclin E adenovirus and fixed 24 h after infection. To induce replication well in triplicate in six-well dishes. The WEE1 inhibitor MK-1775 stress with WEE1 inhibition, cells were treated with 0.25 µM MK-1775 was diluted in growth medium and added directly to the cells for 24 h 18 h after viral infection and then fixed 24 h after infection. At the time before switching to fresh medium. Growth medium was replenished of harvesting, cells were washed three times with PBS and then extracted every 3 d. Colonies containing ∼50 surviving cells were stained with for 5 min in cold CSK buffer and fixed for 10 min in 2% PFA in PBS. 0.05% crystal violet in 1× PBS containing 1% methanol and 1% form- For PCNA staining, cells were fixed with cold methanol for 20 min at aldehyde. The ImageJ plugin ColonyArea (Guzmán et al., 2014) was −20°C instead of PFA. For experiments not requiring antibody staining, used to automatically quantify stained colonies. In some experiments, cells were covered with VECTASHIE LD Antifade Mounting Medium lentivirally expressed sgRNAs were used to functionally inactivate containing DAPI (Vector Laboratories) and imaged within 2 h. For an- PolQ and Brca1 genes exactly as described previously (Wyatt et al., tibody staining, cells were blocked in 3% BSA + 5% normal donkey 2016). In brief, MEFs were infected with lentivirus encoding Cas9 and serum in PBS for 1 h and then incubated in primary antibody for 1 h at either control nontargeting guides or sgRNAs targeting mouse PolQ room temperature. Cells were washed three times with PBS and then and Brca1 genes. Lentivirus-infected cells were selected in puromy- incubated with secondary antibody (1:300) for 1 h at room temperature. cin-containing medium for 3 d. Puromycin-resistant cells were seeded After three washes with PBS, cells were covered with Vectashield Solu- at a density of 1,000 cells/well in triplicate six-well dishes, treated with tion and imaged within 2 h. Antibodies used for fluorescence microscopy MK-1775, and analyzed for clonogenic survival exactly as described in this study were: mouse-RPA32 (1:200; ab2), mouse-G4 (1:200; 1H6), in Figs. 7 D and S4. The PolQ sgRNA was validated as previously de- and rabbit anti–phosphor-H3 (1:400; 06-570) from EMD Millipore, and scribed (Wyatt et al., 2016). To validate the gene editing experiments, rabbit-53BP1 (1:300; sc-22760) and mouse PCNA (1:500; sc-56) from quantitative PCR was used to measure Brca1 mRNA levels in puro- Santa Cruz Biotechnology, Inc. Fixed-cell imaging was performed in the mycin-selected cells. The quantitative PCR primer sequences used in University of North Carolina Microscopy Services Laboratory core fa- these experiments were 5′-CTGCCGTCCAAATTCAAGAAGT-3′ cility using an LSM700 confocal laser scanning microscope (ZEISS) and and 5′-CTTGTGCTTCCCTGTAGGCT-3′ (corresponding with Brca1 the following objective lenses: Plan Apochromat 20×/0.80 differential forward and reverse primers, respectively). As a positive control for interference contrast II/0.8 NA, Plan-NEOFLU AR 40×/1.30 oil/1.3 NA, the quantitative PCR reactions, β-actin mRNA was amplified using and Plan Apochromat 63×/1.4 oil differential interference contrast/1.4 the primers 5′-TTCTTTGCAGCTCCTTCGTTGCCG-3′ (forward) NA. Imaging was performed at 25°C. Flourochromes used were DAPI, and 5′-TGGATGGCTACGTACATGGCTGGG-3′ (reverse). In three Alexa Fluor 488, and Alexa Fluor 647. Images were acquired directly independent experiments, Brca1 mRNA was reduced by 54 ± 4% in using ZEN 2011 Acquisition software (ZEISS). Images were analyzed sgBrca1-transduced MEFs when compared with controls. After lenti- using ImageJ and saved as TIFF files with no further adjustment.

3112 JCB • Volume 216 • Number 10 • 2017 Mitotic spreads which is supported in part by the Cancer Center Core Support grant Mitotic spreads were prepared and analyzed for PCC assays as de- P30 CA016086 to the UNC Lineberger Comprehensive Cancer Center. scribed previously (Nghiem et al., 2001). After a 6-h incubation with The authors declare no competing financial interests. 0.25 µM MK-1775 (or DMSO for controls), cells were harvested with Author contributions: all of the authors designed and discussed trypsin and collected by centrifugation at 300 g for 10 min. Cell pellets experiments. Y. Yang, Y. Gao, M. Durando, W. Ding, and D. Wyatt were resuspended by gentle pipetting in 10 ml of 75 mM KCl, and conducted the experimental work. C. Vaziri conceived the project, su- the resulting suspensions were incubated for 10 min at room tempera- pervised experiments, provided funding, and wrote the manuscript. ture. Cells were then collected by centrifugation (300 g for 10 min) All authors helped with editing. and resuspended in 5 ml of freshly prepared Carnoy’s fixative (three parts methanol and one part glacial acetic acid) and incubated for 10 Submitted: 1 February 2017 min at room temperature. Fixed cells were collected by centrifugation Revised: 27 June 2017 (300 g and 10 min) and resuspended in 200 µl of Carnoy’s fixative. Accepted: 21 July 2017 10 µl of each cell suspension was dropped from a height of 10 cm onto a glass slide and allowed to dry. 15 µl of DAPI solution (Vectashield antifade mounting medium with DAPI fluorochrome) was spotted onto References each slide. Coverslips were gently placed above the DAPI droplet, and Bartkova, J., N. Rezaei, M. Liontos, P. Karakaidos, D. Kletsas, N. Issaeva, edges were sealed with clear nail polish. A BX61 upright widefield L.V. Vassiliou, E. Kolettas, K. Niforou, V.C. Zoumpourlis, et al. 2006. fluorescence microscope (Olympus) was used to identify and count Oncogene-induced senescence is part of the tumorigenesis barrier imposed by DNA damage checkpoints. Nature. 444:633–637. http://dx mitotic cells that had characteristic features of either a normal mitosis .doi.org/10.1038/nature05268 or PCC. Imaging was performed at 25°C using a Plan Apochromat N Bavoux, C., A.M. Leopoldino, V. Bergoglio, J. O-Wang, T. Ogi, A. Bieth, 60×/1.42 oil UIS 2 BFP1 objective lens with an NA of 1.42. Images J.-G. Judde, S.D.J. Pena, M.-F. Poupon, T. Helleday, et al. 2005. Up- were acquired using an ORCA RC camera (Hamamatsu Photonics) regulation of the error-prone DNA polymerase κ promotes pleiotropic genetic alterations and tumorigenesis. Cancer Res. 65:325–330. and Improvision’s Velocity software. Subsequent to acquisition, images Beck, H., V. Nähse, M.S. Larsen, P. Groth, T. Clancy, M. Lees, M. Jørgensen, were analyzed with ImageJ and saved as TIFF files with no additional T. Helleday, R.G. Syljuåsen, and C.S. Sørensen. 2010. Regulators of adjustment. 200 metaphase cells were scored for each independent ex- cyclin-dependent kinases are crucial for maintaining genome integrity periment, and percentages of PCC were calculated using compiled data in S phase. J. Cell Biol. 188:629–638. http://dx.doi.org/10.1083/jcb .200905059 from three separate experiments. Beck, H., V. Nähse-Kumpf, M.S. Larsen, K.A. O’Hanlon, S. Patzke, C. Holmberg, J. Mejlvang, A. Groth, O. Nielsen, R.G. Syljuåsen, and Statistics C.S. Sørensen. 2012. Cyclin-dependent kinase suppression by WEE1 kinase protects the genome through control of replication initiation and Statistical analyses were performed using Prism 6. The unpaired two- nucleotide consumption. Mol. Cell. Biol. 32:4226–4236. http://dx.doi.org tailed Student’s t test was used to compare two datasets in various ex- /10.1128/MCB.00412-12 periments as indicated in the figure legends. In some experiments in Bergoglio, V., A.S. Boyer, E. Walsh, V. Naim, G. Legube, M.Y. Lee, L. Rey, which more than two datasets were compared, we performed ANOVA F. Rosselli, C. Cazaux, K.A. Eckert, and J.S. Hoffmann. 2013. DNA synthesis by Pol η promotes fragile site stability by preventing under- with post hoc Tukey’s honest significance difference test. On graphs replicated DNA in mitosis. J. Cell Biol. 201:395–408. http://dx.doi.org and charts, asterisks indicate p-values: *, P < 0.05; **, P < 0.01; ***, P /10.1083/jcb.201207066 < 0.001. P > 0.05 indicates a difference that is not significant. Bester, A.C., M. Roniger, Y.S. Oren, M.M. Im, D. Sarni, M. Chaoat, A. Bensimon, G. Zamir, D.S. Shewach, and B. Kerem. 2011. Nucleotide deficiency promotes genomic instability in early stages of cancer development. Cell. Online supplemental material 145:435–446. http://dx.doi.org/10.1016/j.cell.2011.03.044 Fig. S1 is related to Fig. 1 and tests the differential effects of various Bétous, R., L. Rey, G. Wang, M.J. Pillaire, N. Puget, J. Selves, D.S. Biard, CDK2 and CDK1 activators (including Cyclins E, A, B, and CDC25C) K. Shin-ya, K.M. Vasquez, C. Cazaux, and J.S. Hoffmann. 2009. Role of TLS DNA polymerases eta and kappa in processing naturally occurring on PCNA monoubiquitination. Fig. S2 shows validation of the flow cy- structured DNA in human cells. Mol. Carcinog. 48:369–378. http://dx.doi tometry–based ssDNA assays used in Figs. 3 and 5. Fig. S3 shows that .org/10.1002/mc.20509 bypass of the G2/M checkpoint by overexpressed CDC25C promotes Bi, X., D.M. Slater, H. Ohmori, and C. Vaziri. 2005. DNA polymerase κ is the accumulation of mitotic cells harboring persistent ssDNA. Fig. S4 specifically required for recovery from the benzo[a]pyrene-dihydrodiol epoxide (BPDE)-induced S-phase checkpoint. J. Biol. Chem. 280:22343– shows that Polq-deficient cells lacking Polθ (the key mediator of the 22355. http://dx.doi.org/10.1074/jbc.M501562200 alternative end-joining DSB repair pathway) are sensitive to MK-1775. Bi, X., L.R. Barkley, D.M. Slater, S. Tateishi, M. Yamaizumi, H. Ohmori, and Fig. S5 shows that Rad18 and Polkκ have separable roles in tolerance C. Vaziri. 2006. Rad18 regulates DNA polymerase κ and is required for recovery from S-phase checkpoint-mediated arrest. Mol. Cell. Biol. of the G4 quadruplex DNA-stabilizing agent PDS. 26:3527–3540. http://dx.doi.org/10.1128/MCB.26.9.3527-3540.2006 Bienko, M., C.M. Green, N. Crosetto, F. Rudolf, G. Zapart, B. Coull, Acknowledgments P. Kannouche, G. Wider, M. Peter, A.R. Lehmann, et al. 2005. Ubiquitin- binding domains in Y-family polymerases regulate translesion synthesis. Science. 310:1821–1824. http://dx.doi.org/10.1126/science.1120615 We thank Dr. Tovah Day for helpful discussions and advice throughout Canman, C.E. 2001. Replication checkpoint: Preventing mitotic catastrophe. Curr. the study. We are grateful to Dr. Wen Huang for providing the R script Biol. 11:R121–R124. http://dx.doi.org/10.1016/S0960-9822(01)00057-4 used to generate the DNA fiber plots. We thank Dr. Rick Wood for Cea, V., L. Cipolla, and S. Sabbioneda. 2015. Replication of structured DNA and providing Polq+/+ and Polq−/− MEFs. its implication in epigenetic stability. Front. Genet. 6:209. http://dx.doi .org/10.3389/fgene.2015.00209 This study was supported by the National Institutes of Health Ceccaldi, R., J.C. Liu, R. Amunugama, I. Hajdu, B. Primack, M.I. Petalcorin, grant R01 ES09558 to C. Vaziri, a University of North Carolina Insti- K.W. O’Connor, P.A. Konstantinopoulos, S.J. Elledge, S.J. Boulton, et tute of Clinical and Translational Sciences Clinical and Translational al. 2015. Homologous-recombination-deficient tumours are dependent Science Award UL1TR001111 to Y. Yang, and a University of North on Polθ-mediated repair. Nature. 518:258–262. http://dx.doi.org/10.1038 /nature14184 Carolina Lineberger Comprehensive Cancer Center postdoctoral Costantino, L., S.K. Sotiriou, J.K. Rantala, S. Magin, E. Mladenov, T. Helleday, award to Y. Gao. Microscopy was performed at the Microscopy Ser- J.E. Haber, G. Iliakis, O.P. Kallioniemi, and T.D. Halazonetis. 2014. vices Laboratory at the University of North Carolina, Chapel Hill, Break-induced replication repair of damaged forks induces genomic

RAD18 confers oncogenic stress tolerance • Yang et al. 3113 duplications in human cells. Science. 343:88–91. http://dx.doi.org/10 Huang, J., M.S. Huen, H. Kim, C.C. Leung, J.N. Glover, X. Yu, and J. Chen. .1126/science.1243211 2009. RAD18 transmits DNA damage signalling to elicit homologous Daigaku, Y., A.A. Davies, and H.D. Ulrich. 2010. Ubiquitin-dependent DNA recombination repair. Nat. Cell Biol. 11:592–603. http://dx.doi.org/10 damage bypass is separable from genome replication. Nature. 465:951– .1038/ncb1865 955. http://dx.doi.org/10.1038/nature09097 Jin, P., S. Hardy, and D.O. Morgan. 1998. Nuclear localization of cyclin B1 Davidson, I.F., A. Li, and J.J. Blow. 2006. Deregulated replication licensing controls mitotic entry after DNA damage. J. Cell Biol. 141:875–885. http causes DNA fragmentation consistent with head-to-tail fork collision. ://dx.doi.org/10.1083/jcb.141.4.875 Mol. Cell. 24:433–443. http://dx.doi.org/10.1016/j.molcel.2006.09.010 Jones, R.M., O. Mortusewicz, I. Afzal, M. Lorvellec, P. García, T. Helleday, and E. Petermann. 2013. Increased replication initiation and conflicts with Davies, A.A., D. Huttner, Y. Daigaku, S. Chen, and H.D. Ulrich. 2008. Activation transcription underlie Cyclin E-induced replication stress. Oncogene. of ubiquitin-dependent DNA damage bypass is mediated by replication 32:3744–3753. http://dx.doi.org/10.1038/onc.2012.387 protein A. Mol. Cell. 29:625–636. http://dx.doi.org/10.1016/j.molcel .2007.12.016 Kannouche, P., B.C. Broughton, M. Volker, F. Hanaoka, L.H. Mullenders, and A.R. Lehmann. 2001. Domain structure, localization, and function of Day, T.A., K. Palle, L.R. Barkley, N. Kakusho, Y. Zou, S. Tateishi, A. Verreault, DNA polymerase , defective in xeroderma pigmentosum variant cells. H. Masai, and C. Vaziri. 2010. Phosphorylated Rad18 directs DNA η Genes Dev. 15:158–172. http://dx.doi.org/10.1101/gad.187501 polymerase η to sites of stalled replication. J. Cell Biol. 191:953–966. http://dx.doi.org/10.1083/jcb.201006043 Lin, J.R., M.K. Zeman, J.Y. Chen, M.C. Yee, and K.A. Cimprich. 2011. SHPRH and HLTF act in a damage-specific manner to coordinate different forms DeNicola, G.M., F.A. Karreth, T.J. Humpton, A. Gopinathan, C. Wei, of postreplication repair and prevent mutagenesis. Mol. Cell. 42:237–249. K. Frese, D. Mangal, K.H. Yu, C.J. Yeo, E.S. Calhoun, et al. 2011. http://dx.doi.org/10.1016/j.molcel.2011.02.026 Oncogene-induced Nrf2 transcription promotes ROS detoxification and tumorigenesis. Nature. 475:106–109. http://dx.doi.org/10.1038/ Liontos, M., M. Koutsami, M. Sideridou, K. Evangelou, D. Kletsas, B. Levy, nature10189 A. Kotsinas, O. Nahum, V. Zoumpourlis, M. Kouloukoussa, et al. 2007. Deregulated overexpression of hCdt1 and hCdc6 promotes malignant Despras, E., M. Sittewelle, C. Pouvelle, N. Delrieu, A.M. Cordonnier, and behavior. Cancer Res. 67:10899–10909. http://dx.doi.org/10.1158/0008 P.L. Kannouche. 2016. Rad18-dependent SUMOylation of human -5472.CAN-07-2837 specialized DNA polymerase eta is required to prevent under-replicated DNA. Nat. Commun. 7:13326. http://dx.doi.org/10.1038/ncomms13326 Liu, P., L.R. Barkley, T. Day, X. Bi, D.M. Slater, M.G. Alexandrow, H.P. Nasheuer, and C. Vaziri. 2006. The Chk1-mediated S-phase checkpoint targets Diamant, N., A. Hendel, I. Vered, T. Carell, T. Reissner, N. de Wind, N. Geacinov, initiation factor Cdc45 via a Cdc25A/Cdk2-independent mechanism. and Z. Livneh. 2012. DNA damage bypass operates in the S and G2 J. Biol. Chem. 281:30631–30644. http://dx.doi.org/10.1074/jbc phases of the cell cycle and exhibits differential mutagenicity. Nucleic .M602982200 Acids Res. 40:170–180. http://dx.doi.org/10.1093/nar/gkr596 Lopes, M., M. Foiani, and J.M. Sogo. 2006. Multiple mechanisms control Di Micco, R., M. Fumagalli, A. Cicalese, S. Piccinin, P. Gasparini, C. Luise, chromosome integrity after replication fork uncoupling and restart at C. Schurra, M. Garre’, P.G. Nuciforo, A. Bensimon, et al. 2006. irreparable UV lesions. Mol. Cell. 21:15–27. http://dx.doi.org/10.1016/j Oncogene-induced senescence is a DNA damage response triggered by .molcel.2005.11.015 DNA hyper-replication. Nature. 444:638–642. http://dx.doi.org/10.1038 /nature05327 Maher, V.M., L.M. Ouellette, R.D. Curren, and J.J. McCormick. 1976. Caffeine enhancement of the cytotoxic and mutagenic effect of ultraviolet Durando, M., S. Tateishi, and C. Vaziri. 2013. A non-catalytic role of irradiation in a xeroderma pigmentosum variant strain of human cells. DNA polymerase η in recruiting Rad18 and promoting PCNA Biochem. Biophys. Res. Commun. 71:228–234. http://dx.doi.org/10.1016 monoubiquitination at stalled replication forks. Nucleic Acids Res. /0006-291X(76)90272-2 41:3079–3093. http://dx.doi.org/10.1093/nar/gkt016 Meerbrey, K.L., G. Hu, J.D. Kessler, K. Roarty, M.Z. Li, J.E. Fang, Eddy, S., M. Tillman, L. Maddukuri, A. Ketkar, M.K. Zafar, and R.L. Eoff. J.I. Herschkowitz, A.E. Burrows, A. Ciccia, T. Sun, et al. 2011. The pIND 2016. Human translesion polymerase κ exhibits enhanced activity and UCER lentiviral toolkit for inducible RNA interference in vitro and in reduced fidelity two nucleotides from G-quadruplex DNA. Biochemistry. vivo. Proc. Natl. Acad. Sci. USA. 108:3665–3670. http://dx.doi.org/10 55:5218–5229. http://dx.doi.org/10.1021/acs.biochem.6b00374 .1073/pnas.1019736108 Fikaris, A.J., A.E. Lewis, A. Abulaiti, O.M. Tsygankova, and J.L. Meinkoth. Neelsen, K.J., I.M. Zanini, R. Herrador, and M. Lopes. 2013a. Oncogenes 2006. Ras triggers ataxia-telangiectasia-mutated and Rad-3-related induce genotoxic stress by mitotic processing of unusual replication activation and apoptosis through sustained mitogenic signaling. J. Biol. intermediates. J. Cell Biol. 200:699–708. http://dx.doi.org/10.1083/jcb Chem. 281:34759–34767. http://dx.doi.org/10.1074/jbc.M606737200 .201212058 Gao, Y., E. Mutter-Rottmayer, A.M. Greenwalt, D. Goldfarb, F. Yan, Y. Yang, Neelsen, K.J., I.M. Zanini, S. Mijic, R. Herrador, R. Zellweger, A. Ray R.C. Martinez-Chacin, K.H. Pearce, S. Tateishi, M.B. Major, and Chaudhuri, K.D. Creavin, J.J. Blow, and M. Lopes. 2013b. Deregulated C. Vaziri. 2016a. A neomorphic cancer cell-specific role of MAGE-A4 in origin licensing leads to chromosomal breaks by rereplication of a gapped trans-lesion synthesis. Nat. Commun. 7:12105. http://dx.doi.org/10.1038 DNA template. Genes Dev. 27:2537–2542. http://dx.doi.org/10.1101/gad /ncomms12105 .226373.113 Gao, Y., S. Tateishi, and C. Vaziri. 2016b. Pathological trans-lesion synthesis in Nghiem, P., P.K. Park, Y. Kim, C. Vaziri, and S.L. Schreiber. 2001. ATR inhibition cancer. Cell Cycle. 15:3005–3006. http://dx.doi.org/10.1080/15384101 selectively sensitizes G1 checkpoint-deficient cells to lethal premature .2016.1214045 chromatin condensation. Proc. Natl. Acad. Sci. USA. 98:9092–9097. http Garcia-Exposito, L., E. Bournique, V. Bergoglio, A. Bose, J. Barroso-Gonzalez, ://dx.doi.org/10.1073/pnas.161281798 S. Zhang, J.L. Roncaioli, M. Lee, C.T. Wallace, S.C. Watkins, et al. Olive, P.L., J.P. Banáth, and R.E. Durand. 1990. Heterogeneity in radiation- 2016. Proteomic profiling reveals a specific role for translesion DNA induced DNA damage and repair in tumor and normal cells measured polymerase η in the alternative lengthening of telomeres. Cell Reports. using the “comet” assay. Radiat. Res. 122:86–94. http://dx.doi.org/10 17:1858–1871. http://dx.doi.org/10.1016/j.celrep.2016.10.048 .2307/3577587 Guzmán, C., M. Bagga, A. Kaur, J. Westermarck, and D. Abankwa. 2014. Palle, K., and C. Vaziri. 2011. Rad18 E3 ubiquitin ligase activity mediates ColonyArea: An ImageJ plugin to automatically quantify colony Fanconi anemia pathway activation and cell survival following DNA formation in clonogenic assays. PLoS One. 9:e92444. http://dx.doi.org Topoisomerase 1 inhibition. Cell Cycle. 10:1625–1638. http://dx.doi.org /10.1371/journal.pone.0092444 /10.4161/cc.10.10.15617 Hahn, W.C., C.M. Counter, A.S. Lundberg, R.L. Beijersbergen, M.W. Brooks, Prakash, S., R.E. Johnson, and L. Prakash. 2005. Eukaryotic translesion and R.A. Weinberg. 1999. Creation of human tumour cells with defined synthesis DNA polymerases: Specificity of structure and function. Annu. genetic elements. Nature. 400:464–468. http://dx.doi.org/10.1038/22780 Rev. Biochem. 74:317–353. http://dx.doi.org/10.1146/annurev.biochem Halazonetis, T.D., V.G. Gorgoulis, and J. Bartek. 2008. An oncogene-induced .74.082803.133250 DNA damage model for cancer development. Science. 319:1352–1355. Raderschall, E., E.I. Golub, and T. Haaf. 1999. Nuclear foci of mammalian http://dx.doi.org/10.1126/science.1140735 recombination proteins are located at single-stranded DNA regions Hedglin, M., and S.J. Benkovic. 2015. Regulation of Rad6/Rad18 activity during formed after DNA damage. Proc. Natl. Acad. Sci. USA. 96:1921–1926. DNA damage tolerance. Annu. Rev. Biophys. 44:207–228. http://dx.doi http://dx.doi.org/10.1073/pnas.96.5.1921 .org/10.1146/annurev-biophys-060414-033841 Räschle, M., G. Smeenk, R.K. Hansen, T. Temu, Y. Oka, M.Y. Hein, N. Nagaraj, Hishiki, A., H. Hashimoto, T. Hanafusa, K. Kamei, E. Ohashi, T. Shimizu, D.T. Long, J.C. Walter, K. Hofmann, et al. 2015. Proteomics reveals H. Ohmori, and M. Sato. 2009. Structural basis for novel interactions dynamic assembly of repair complexes during bypass of DNA cross- between human translesion synthesis polymerases and proliferating cell links. Science. 348:1253671. http://dx.doi.org/10.1126/science.1253671 nuclear antigen. J. Biol. Chem. 284:10552–10560. http://dx.doi.org/10 Regairaz, M., Y.W. Zhang, H. Fu, K.K. Agama, N. Tata, S. Agrawal, .1074/jbc.M809745200 M.I. Aladjem, and Y. Pommier. 2011. Mus81-mediated DNA cleavage

3114 JCB • Volume 216 • Number 10 • 2017 resolves replication forks stalled by topoisomerase I-DNA complexes. Vaziri, C., S. Saxena, Y. Jeon, C. Lee, K. Murata, Y. Machida, N. Wagle, J. Cell Biol. 195:739–749. http://dx.doi.org/10.1083/jcb.201104003 D.S. Hwang, and A. Dutta. 2003. A p53-dependent checkpoint pathway prevents rereplication. Mol. Cell. 11:997–1008. http://dx.doi.org/10.1016 Rosenblatt, J., Y. Gu, and D.O. Morgan. 1992. Human cyclin-dependent kinase /S1097-2765(03)00099-6 2 is activated during the S and G2 phases of the cell cycle and associates with cyclin A. Proc. Natl. Acad. Sci. USA. 89:2824–2828. http://dx.doi Volpe, J.P., and J.E. Cleaver. 1995. Xeroderma pigmentosum variant cells are .org/10.1073/pnas.89.7.2824 resistant to immortalization. Mutat. Res. 337:111–117. http://dx.doi.org /10.1016/0921-8777(95)00015-C Sakurikar, N., R. Thompson, R. Montano, and A. Eastman. 2016. A subset of cancer cell lines is acutely sensitive to the Chk1 inhibitor MK-8776 as Watanabe, K., S. Tateishi, M. Kawasuji, T. Tsurimoto, H. Inoue, and monotherapy due to CDK2 activation in S phase. Oncotarget. 7:1380– M. Yamaizumi. 2004. Rad18 guides poleta to replication stalling sites 1394. http://dx.doi.org/10.18632/oncotarget.6364 through physical interaction and PCNA monoubiquitination. EMBO J. 23:3886–3896. http://dx.doi.org/10.1038/sj.emboj.7600383 Schoppy, D.W., R.L. Ragland, O. Gilad, N. Shastri, A.A. Peters, M. Murga, Williams, S.A., S. Longerich, P. Sung, C. Vaziri, and G.M. Kupfer. 2011. The E3 O. Fernandez-Capetillo, J.A. Diehl, and E.J. Brown. 2012. Oncogenic ubiquitin ligase RAD18 regulates ubiquitylation and chromatin loading stress sensitizes murine cancers to hypomorphic suppression of ATR. of FANCD2 and FANCI. Blood. 117:5078–5087. http://dx.doi.org/10 J. Clin. Invest. 122:241–252. http://dx.doi.org/10.1172/JCI58928 .1182/blood-2010-10-311761 Sogo, J.M., M. Lopes, and M. Foiani. 2002. Fork reversal and ssDNA Wyatt, D.W., W. Feng, M.P. Conlin, M.J. Yousefzadeh, S.A. Roberts, accumulation at stalled replication forks owing to checkpoint defects. P. Mieczkowski, R.D. Wood, G.P. Gupta, and D.A. Ramsden. 2016. Science. 297:599–602. http://dx.doi.org/10.1126/science.1074023 Essential roles for polymerase θ-mediated end joining in the repair of Sørensen, C.S., and R.G. Syljuåsen. 2012. Safeguarding genome integrity: the chromosome breaks. Mol. Cell. 63:662–673. http://dx.doi.org/10.1016/j checkpoint kinases ATR, CHK1 and WEE1 restrain CDK activity during .molcel.2016.06.020 normal DNA replication. Nucleic Acids Res. 40:477–486. http://dx.doi Yanagihara, H., J. Kobayashi, S. Tateishi, A. Kato, S. Matsuura, H. Tauchi, .org/10.1093/nar/gkr697 K. Yamada, J. Takezawa, K. Sugasawa, C. Masutani, et al. 2011. NBS1 Soria, G., O. Podhajcer, C. Prives, and V. Gottifredi. 2006. P21Cip1/WAF1 recruits RAD18 via a RAD6-like domain and regulates Pol η-dependent downregulation is required for efficient PCNA ubiquitination after UV translesion DNA synthesis. Mol. Cell. 43:788–797. http://dx.doi.org/10 irradiation. Oncogene. 25:2829–2838. http://dx.doi.org/10.1038/sj.onc .1016/j.molcel.2011.07.026 .1209315 Yang, Y., M. Durando, S.L. Smith-Roe, C. Sproul, A.M. Greenwalt, W. Kaufmann, S. Oh, E.A. Hendrickson, and C. Vaziri. 2013. Cell cycle Srinivasan, S.V., D. Dominguez-Sola, L.C. Wang, O. Hyrien, and J. Gautier. stage-specific roles of Rad18 in tolerance and repair of oxidative DNA 2013. Cdc45 is a critical effector of myc-dependent DNA replication damage. Nucleic Acids Res. 41:2296–2312. http://dx.doi.org/10.1093/nar stress. Cell Reports. 3:1629–1639. http://dx.doi.org/10.1016/j.celrep /gks1325 .2013.04.002 Yang, Y., J.C. Poe, L. Yang, A. Fedoriw, S. Desai, T. Magnuson, Z. Li, Terai, K., T. Abbas, A.A. Jazaeri, and A. Dutta. 2010. CRL4(Cdt2) E3 ubiquitin Y. Fedoriw, K. Araki, Y. Gao, et al. 2016. Rad18 confers hematopoietic ligase monoubiquitinates PCNA to promote translesion DNA synthesis. progenitor cell DNA damage tolerance independently of the Fanconi Mol. Cell. 37:143–149. http://dx.doi.org/10.1016/j.molcel.2009.12.018 Anemia pathway in vivo. Nucleic Acids Res. 44:4174–4188. http://dx.doi Tsuji, Y., K. Watanabe, K. Araki, M. Shinohara, Y. Yamagata, T. Tsurimoto, .org/10.1093/nar/gkw072 F. Hanaoka, K. Yamamura, M. Yamaizumi, and S. Tateishi. 2008. Zlatanou, A., E. Despras, T. Braz-Petta, I. Boubakour-Azzouz, C. Pouvelle, Recognition of forked and single-stranded DNA structures by human G.S. Stewart, S. Nakajima, A. Yasui, A.A. Ishchenko, and P.L. Kannouche. RAD18 complexed with RAD6B protein triggers its recruitment to 2011. The hMsh2-hMsh6 complex acts in concert with monoubiquitinated stalled replication forks. Genes Cells. 13:343–354. http://dx.doi.org/10 PCNA and Pol η in response to oxidative DNA damage in human cells. .1111/j.1365-2443.2008.01176.x Mol. Cell. 43:649–662. http://dx.doi.org/10.1016/j.molcel.2011.06.023

RAD18 confers oncogenic stress tolerance • Yang et al. 3115